[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN116664213A - Agricultural product market price early warning management system and method based on big data analysis - Google Patents

Agricultural product market price early warning management system and method based on big data analysis Download PDF

Info

Publication number
CN116664213A
CN116664213A CN202310637249.2A CN202310637249A CN116664213A CN 116664213 A CN116664213 A CN 116664213A CN 202310637249 A CN202310637249 A CN 202310637249A CN 116664213 A CN116664213 A CN 116664213A
Authority
CN
China
Prior art keywords
training
time sequence
data
feature
market transaction
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202310637249.2A
Other languages
Chinese (zh)
Inventor
杨景榆
湛碧琨
张峻华
庾莎菲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dongguan Zhuocheng Culture Development Co ltd
Original Assignee
Dongguan Zhuocheng Culture Development Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dongguan Zhuocheng Culture Development Co ltd filed Critical Dongguan Zhuocheng Culture Development Co ltd
Priority to CN202310637249.2A priority Critical patent/CN116664213A/en
Publication of CN116664213A publication Critical patent/CN116664213A/en
Priority to CN202311247179.6A priority patent/CN117094782A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0283—Price estimation or determination
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Forestry; Mining
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Evolutionary Biology (AREA)
  • Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Marketing (AREA)
  • Animal Husbandry (AREA)
  • Biomedical Technology (AREA)
  • Primary Health Care (AREA)
  • Human Resources & Organizations (AREA)
  • Mining & Mineral Resources (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Agronomy & Crop Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Tourism & Hospitality (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application relates to the technical field of market management, and particularly discloses an agricultural product market price early warning management system and method based on big data analysis.

Description

Agricultural product market price early warning management system and method based on big data analysis
Technical Field
The application relates to the technical field of market management, in particular to an agricultural product market price early warning management system and method based on big data analysis.
Background
Agricultural products are key industries for national agricultural economic development, and have important significance for guaranteeing national grain safety and livelihood of farmers. However, the existing agricultural product market price analysis method mostly adopts a qualitative prediction method, namely, analysis is carried out by experience of expert staff, and the problem that the experience richness of the expert staff is different exists, so that an analysis result is subjective, and unified and standard standardized agricultural product market price analysis cannot be realized. Meanwhile, the market price of the agricultural product cannot be accurately analyzed through various factors by means of the experience of expert personnel, so that the accuracy of the market price prediction result of the agricultural product is low, early warning of the fluctuation of the price of the agricultural product cannot be timely and accurately performed, unbalance of supply and demand of the market of the agricultural product is caused, and great influence is brought to production operators and consumers of the market of the agricultural product.
Therefore, an optimized agricultural product market price pre-warning management system based on big data analysis is desired.
Disclosure of Invention
The application provides an agricultural product market price early warning management system and method based on big data analysis, which constructs an agricultural product market price prediction model based on market transaction data and meteorological data of agricultural products to be predicted so as to generate agricultural product market price early warning information, in particular, the system and the method acquire time sequence dynamic change characteristics of the meteorological data and full expression of time sequence collaborative correlation characteristic information between the time sequence dynamic change characteristics of the market transaction data of the agricultural products to be predicted through an artificial intelligence technology, so that price prediction of the agricultural products is accurately carried out, price early warning is carried out on the agricultural products with excessive price, market price stability of the agricultural products is further ensured, and powerful support is provided for macroscopic regulation and farmer specified production strategies by government.
In a first aspect, an early warning management system for market price of agricultural products based on big data analysis is provided, the early warning management system for market price of agricultural products based on big data analysis includes: the data acquisition module is used for acquiring meteorological data and market transaction data of agricultural products to be predicted for a plurality of days in a preset time period; the independent heat coding module is used for respectively carrying out independent heat coding on the meteorological data of each day to obtain a plurality of meteorological data independent heat coding vectors; the meteorological time sequence associated feature extraction module is used for arranging the plurality of meteorological data single-heat coding vectors into one-dimensional input vectors and then obtaining meteorological time sequence feature vectors through a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model; the market transaction time sequence correlation feature extraction module is used for arranging market transaction data of agricultural products to be predicted for a plurality of days in the preset time period into a market transaction data time sequence input vector according to a time dimension, and obtaining a market transaction time sequence feature vector through a market transaction data time sequence feature extractor based on a one-dimensional convolutional neural network model; the feature fusion module is used for fusing the meteorological time sequence feature vector and the market transaction time sequence feature vector by using a cascading function to obtain a decoding feature vector; the price prediction module is used for carrying out decoding regression on the decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is used for representing a price prediction value of the agricultural product to be predicted; and the price early warning module is used for determining whether to generate price early warning prompt information or not based on the comparison between the decoding value and a preset threshold value.
With reference to the first aspect, in an implementation manner of the first aspect, the method further includes a training module for training the weather timing feature extractor based on the one-dimensional convolutional neural network model, the market transaction data timing feature extractor based on the one-dimensional convolutional neural network model, and the decoder.
With reference to the first aspect, in an implementation manner of the first aspect, the training module includes: the system comprises a training data acquisition module, a data processing module and a data processing module, wherein the training data acquisition module is used for acquiring training data, the training data comprises training meteorological data for a plurality of days in a preset time period, training market transaction data of agricultural products to be predicted and a true value of a price prediction value of the agricultural products to be predicted; the training independent heat coding module is used for independent heat coding the training meteorological data of each day to obtain a plurality of training meteorological data independent heat coding vectors; the training weather time sequence associated feature extraction module is used for arranging the plurality of training weather data single-heat coding vectors into training one-dimensional input vectors and then obtaining training weather time sequence feature vectors through the weather time sequence feature extractor based on the one-dimensional convolutional neural network model; the training market transaction time sequence correlation feature extraction module is used for arranging training market transaction data of agricultural products to be predicted for a plurality of days in the preset time period into training market transaction data time sequence input vectors according to time dimension, and then obtaining training market transaction time sequence feature vectors through the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model; the training feature fusion module is used for fusing the training meteorological time sequence feature vector and the training market transaction time sequence feature vector by using a cascading function to obtain a training decoding feature vector; a decoding loss module, configured to pass the training decoding feature vector through the decoder to obtain a decoding loss function value; the model training module is used for training the weather time sequence feature extractor based on the one-dimensional convolutional neural network model, the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model and the decoder based on the decoding loss function value and through gradient descent direction propagation, wherein in each round of iteration of training, a weight matrix of the decoder is subjected to half-space structuring constraint iteration of weight intrinsic support.
In a second aspect, there is provided a method for early warning management of market price of agricultural products based on big data analysis, the method comprising: acquiring meteorological data and market transaction data of agricultural products to be predicted for a plurality of days within a preset time period; the weather data of each day are subjected to independent heat coding respectively to obtain a plurality of weather data independent heat coding vectors; the plurality of meteorological data independent-heat coding vectors are arranged into one-dimensional input vectors and then a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model is used for obtaining meteorological time sequence feature vectors; the market transaction data of the agricultural products to be predicted for a plurality of days in the preset time period are arranged into a time sequence input vector of the market transaction data according to the time dimension, and then the time sequence input vector of the market transaction data is obtained through a time sequence feature extractor of the market transaction data based on a one-dimensional convolutional neural network model; fusing the weather timing feature vector and the market transaction timing feature vector using a cascading function to obtain a decoded feature vector; carrying out decoding regression on the decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is used for representing a price prediction value of the agricultural product to be predicted; and determining whether to generate price pre-warning prompt information based on a comparison between the decoded value and a predetermined threshold.
In a third aspect, there is provided a chip comprising an input-output interface, at least one processor, at least one memory and a bus, the at least one memory to store instructions, the at least one processor to invoke the instructions in the at least one memory to perform the method in the second aspect.
In a fourth aspect, a computer readable medium is provided for storing a computer program comprising instructions for performing the method of the second aspect described above.
In a fifth aspect, there is provided a computer program product comprising instructions which, when executed by a computer, perform the method of the second aspect described above.
The application provides an agricultural product market price early warning management system and method based on big data analysis, which constructs an agricultural product market price prediction model based on market transaction data and meteorological data of agricultural products to be predicted so as to generate agricultural product market price early warning information, in particular, the system and the method acquire time sequence dynamic change characteristics of the meteorological data and time sequence cooperative correlation characteristic information between the time sequence dynamic change characteristics of the market transaction data of the agricultural products to be predicted through an artificial intelligence technology, so that price prediction of the agricultural products is accurately carried out, price early warning is carried out on the agricultural products with excessive price, market price stability of the agricultural products is further ensured, and powerful support is provided for macroscopic regulation and control of government and appointed production strategies of farmers.
Drawings
FIG. 1 is a schematic block diagram of an agricultural product market price pre-warning management system based on big data analysis in accordance with an embodiment of the present application.
Fig. 2 is a schematic structural diagram of a training module in the agricultural product market price early warning management system based on big data analysis according to the embodiment of the application.
Fig. 3 is a schematic flow chart of an agricultural product market price early warning management method based on big data analysis according to an embodiment of the present application.
Fig. 4 is a schematic diagram of a model architecture of an agricultural product market price early warning management method based on big data analysis according to an embodiment of the present application.
Fig. 5 is a schematic flow chart of a training phase in an agricultural product market price early warning management method based on big data analysis according to an embodiment of the present application.
Fig. 6 is a schematic diagram of a model architecture of a training stage in the agricultural product market price early warning management method based on big data analysis according to the embodiment of the application.
Detailed Description
The technical scheme of the application will be described below with reference to the accompanying drawings.
Because of the deep learning-based deep neural network model, related terms and concepts of the deep neural network model that may be related to embodiments of the present application are described below.
In the deep neural network model, the hidden layers may be convolutional layers and pooled layers. The set of weight values corresponding to the convolutional layer is referred to as a filter, also referred to as a convolutional kernel. The filter and the input eigenvalue are both represented as a multi-dimensional matrix, correspondingly, the filter represented as a multi-dimensional matrix is also called a filter matrix, the input eigenvalue represented as a multi-dimensional matrix is also called an input eigenvalue, of course, besides the input eigenvalue, the eigenvector can also be input, and the input eigenvector is only exemplified by the input eigenvector. The operation of the convolution layer is called a convolution operation, which is to perform an inner product operation on a part of eigenvalues of the input eigenvalue matrix and weight values of the filter matrix.
The operation process of each convolution layer in the deep neural network model can be programmed into software, and then the output result of each layer of network, namely the output characteristic matrix, is obtained by running the software in an operation device. For example, the software performs inner product operation by taking the upper left corner of the input feature matrix of each layer of network as a starting point and taking the size of the filter as a window in a sliding window mode, and extracting data of one window from the feature value matrix each time. After the inner product operation is completed between the data of the right lower corner window of the input feature matrix and the filter, a two-dimensional output feature matrix of each layer of network can be obtained. The software repeats the above process until the entire output feature matrix for each layer of network is generated.
The convolution layer operation process is to slide a window with a filter size across the whole input image (i.e. the input feature matrix), and at each moment, to perform inner product operation on the input feature value covered in the window and the filter, wherein the step length of window sliding is 1. Specifically, the upper left corner of the input feature matrix is used as a starting point, the size of the filter is used as a window, the sliding step length of the window is 1, the input feature value of one window is extracted from the feature value matrix each time and the filter performs inner product operation, and when the data of the lower right corner of the input feature matrix and the filter complete inner product operation, a two-dimensional output feature matrix of the input feature matrix can be obtained.
Since it is often necessary to reduce the number of training parameters, the convolutional layer often requires a periodic introduction of a pooling layer, the only purpose of which is to reduce the spatial size of the image during image processing. The pooling layer may include an average pooling operator and/or a maximum pooling operator for sampling the input image to obtain a smaller size image. The average pooling operator may calculate pixel values in the image over a particular range to produce an average as a result of the average pooling. The max pooling operator may take the pixel with the largest value in a particular range as the result of max pooling. In addition, just as the size of the weighting matrix used in the convolutional layer should be related to the image size, the operators in the pooling layer should also be related to the image size. The size of the image output after the processing by the pooling layer can be smaller than the size of the image input to the pooling layer, and each pixel point in the image output by the pooling layer represents the average value or the maximum value of the corresponding sub-region of the image input to the pooling layer.
Since the functions actually required to be simulated in the deep neural network are nonlinear, but the previous rolling and pooling can only simulate linear functions, in order to introduce nonlinear factors in the deep neural network model to increase the characterization capacity of the whole network, an activation layer is further arranged after the pooling layer, an activation function is arranged in the activation layer, and the commonly used excitation functions include sigmoid, tanh, reLU functions and the like.
As described above, the existing agricultural product market price analysis methods mostly adopt a qualitative prediction method, namely, analysis is performed by experience of expert personnel, and the problem that the experience richness of the expert personnel is different exists, so that analysis results are subjective, and unified and standard standardized agricultural product market price analysis cannot be realized. Meanwhile, the market price of the agricultural product cannot be accurately analyzed through various factors by means of the experience of expert personnel, so that the accuracy of the market price prediction result of the agricultural product is low, early warning of the fluctuation of the price of the agricultural product cannot be timely and accurately performed, unbalance of supply and demand of the market of the agricultural product is caused, and great influence is brought to production operators and consumers of the market of the agricultural product. Therefore, an optimized agricultural product market price pre-warning management system based on big data analysis is desired.
Accordingly, considering that in the actual process of detecting and early warning the market price of the agricultural product, the market price of the agricultural product is closely related to the meteorological condition, in order to realize unified and standardized analysis of the market price of the agricultural product, thereby providing powerful support for the government to macroscopically regulate and control and the farmer to specify the production strategy, in the technical scheme of the application, the agricultural product market price prediction model is expected to be constructed based on the market transaction data and the meteorological data of the agricultural product to be predicted so as to generate the early warning information of the market price of the agricultural product. However, since the meteorological data and the market transaction data of the agricultural product to be predicted have dynamic change rules in the time dimension, and a time sequence cooperative association relationship is also formed between the meteorological data and the market transaction data of the agricultural product to be predicted. Therefore, in the process, the difficulty is how to fully express the time sequence collaborative correlation characteristic information between the time sequence dynamic change characteristic of the meteorological data and the time sequence dynamic change characteristic of the market transaction data of the agricultural products to be predicted, so that the price prediction of the agricultural products is accurately performed, and the price early warning is performed on the agricultural products with excessive price, thereby ensuring the market price stability of the agricultural products, and providing powerful support for the government to perform macroscopic regulation and control and the farmer to specify the production strategy.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. The development of deep learning and neural networks provides new solutions and schemes for mining time sequence collaborative correlation characteristic information between time sequence dynamic change characteristics of meteorological data and time sequence dynamic change characteristics of market transaction data of agricultural products to be predicted.
Specifically, in the technical scheme of the application, firstly, meteorological data and market transaction data of agricultural products to be predicted are obtained for a plurality of days in a preset time period. It should be appreciated that the weather data will typically include a plurality of parameter information such as temperature, humidity, rainfall, etc. Therefore, in order to take the parameter information as the input of the model to comprehensively analyze the time sequence dynamic change rule of the meteorological data, in the technical scheme of the application, the independent heat encoding is further carried out on the meteorological data of each day to obtain a plurality of independent heat encoding vectors of the meteorological data. In particular, here, the use of the single thermal code enables each parameter information in the meteorological data to be a binary vector, facilitating calculation and processing of the model.
Then, considering that each parameter information in the meteorological data has a dynamic time sequence cooperative change rule in a time dimension, in order to fully express the time sequence change characteristics of the meteorological data, the plurality of meteorological data single-heat coding vectors need to be arranged into one-dimensional input vectors and then are coded in a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model, so that time sequence related feature information of the meteorological data is extracted, and then the meteorological time sequence feature vectors are obtained.
Further, for the market transaction data of the agricultural products to be predicted, the market transaction data also has time sequence dynamic change characteristic information in the time dimension, so in the technical scheme of the application, the market transaction data of the agricultural products to be predicted for a plurality of days in the preset time period are further arranged into time sequence input vectors of the market transaction data according to the time dimension, and then the time sequence associated characteristic information of the market transaction data of the agricultural products to be predicted in the time dimension is extracted by the time sequence characteristic extractor of the market transaction data based on the one-dimensional convolutional neural network model, so that the time sequence associated characteristic vectors of the market transaction data of the agricultural products to be predicted are obtained.
In order to integrate the time sequence change characteristics of the market transaction data and the gas phase data of the agricultural products to be predicted to predict the price of the agricultural products, in the technical scheme of the application, a cascading function is further used for fusing the weather time sequence characteristic vector and the market transaction time sequence characteristic vector to obtain a decoding characteristic vector so as to represent time sequence collaborative correlation characteristic information between the time sequence change characteristics of the market transaction data and the weather data of the agricultural products to be predicted. It should be appreciated that the cascading functions enable the network to have a certain logic reasoning capability, so that the association information between the data can be mined. Therefore, in the technical scheme of the application, the time sequence collaborative association change characteristic information between the meteorological data and the market transaction data can be mined by using the cascading function, so that the subsequent more accurate prediction of the price of agricultural products is facilitated.
And then, carrying out decoding regression on the decoding eigenvector through a decoder to obtain a decoding value for representing the price predicted value of the agricultural product to be predicted. That is, the market transaction data of the agricultural product and the time sequence cooperative associated change characteristic information of the meteorological data are decoded so as to predict the price of the agricultural product, and whether to generate the price early warning prompt information is determined based on the comparison between the price prediction value and a predetermined threshold value. Accordingly, in one specific example of the present application, price pre-warning cue information is generated in response to the decoded value being greater than the predetermined threshold.
In particular, in the technical solution of the present application, considering that the weather timing feature vector and the market transaction timing feature vector respectively express the independent thermal coding of the weather data and the time sequence local correlation feature of the market transaction data, when the weather timing feature vector and the market transaction timing feature vector are fused by using a cascade function, the fitting difference, such as over fitting or under fitting, of the weather timing feature vector and the market transaction timing feature vector respectively, may be caused by the point convolution operation, so that the feature distribution of each of the weather timing feature vector and the market transaction timing feature vector to be subjected to the cascade operation has different weight fitting directions with respect to the corresponding part of the weight matrix of the decoder, and thus, the overall feature distribution of the decoded feature vector obtained after the cascade has a problem of poor convergence with respect to the weight matrix of the decoder, thereby affecting the training speed of the decoder.
Based on this, the applicant of the present application performs, during the training process of the decoding eigenvectors, for example denoted as V, a half-space structuring constraint of the weight matrix M for weight eigen support during each iteration of the decoder, for example denoted as M, specifically expressed as:
Wherein V is e Is a matrix M T Eigenvalues of M constitute eigenvector sets.
Here, the weight eigen-supported half-space structuring constraint supports correlation integration of eigenvalue sets of a structuring matrix of a weight matrix M of the decoder and a decoding eigenvector V to be decoded, and performs structural support constraint of a hyperplane as a decision boundary on a half-space (half-space) for coupling with a high-dimensional manifold of the decoding eigenvector V represented by the weight matrix M, so that the high-dimensional manifold of the decoding eigenvector V to be decoded can be effectively converged with respect to the hyperplane in a half-space open domain represented by the weight matrix M, thereby improving training speed of the decoder. Therefore, the price prediction of the agricultural products can be accurately performed, and the price early warning is performed on the agricultural products with excessive price, so that the market price stability of the agricultural products is ensured, and powerful support is provided for the macroscopic regulation and control of the government and the appointed production strategy of farmers.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
FIG. 1 is a schematic block diagram of an agricultural product market price pre-warning management system based on big data analysis in accordance with an embodiment of the present application. As shown in fig. 1, the agricultural product market price early warning management system 100 based on big data analysis includes:
The data acquisition module 110 is used for acquiring meteorological data and market transaction data of agricultural products to be predicted for a plurality of days within a preset time period. It should be understood that, considering that in the actual process of detecting and early warning the market price of the agricultural product, since the market price of the agricultural product is closely related to the meteorological condition, in order to realize unified standardized analysis of the market price of the agricultural product, thereby providing powerful support for the government to perform macroscopic regulation and control and the farmer to specify the production strategy, in the technical scheme of the application, it is desirable to construct the market price prediction model of the agricultural product based on the market transaction data and the meteorological data of the agricultural product to be predicted so as to generate the early warning information of the market price of the agricultural product. However, since the meteorological data and the market transaction data of the agricultural product to be predicted have dynamic change rules in the time dimension, and a time sequence cooperative association relationship is also formed between the meteorological data and the market transaction data of the agricultural product to be predicted. Therefore, in the process, the difficulty is how to fully express the time sequence collaborative correlation characteristic information between the time sequence dynamic change characteristic of the meteorological data and the time sequence dynamic change characteristic of the market transaction data of the agricultural products to be predicted, so that the price prediction of the agricultural products is accurately performed, and the price early warning is performed on the agricultural products with excessive price, thereby ensuring the market price stability of the agricultural products, and providing powerful support for the government to perform macroscopic regulation and control and the farmer to specify the production strategy.
And the independent heat encoding module 120 is configured to independent heat encode the weather data of each day to obtain a plurality of independent heat encoding vectors of the weather data. It will be appreciated that for such weather data, it will typically contain a number of parameter information such as temperature, humidity, rainfall etc. Therefore, in order to take the parameter information as the input of the model to comprehensively analyze the time sequence dynamic change rule of the meteorological data, in the technical scheme of the application, the independent heat encoding is further carried out on the meteorological data of each day to obtain a plurality of independent heat encoding vectors of the meteorological data. In particular, here, the use of the single thermal code enables each parameter information in the meteorological data to be a binary vector, facilitating calculation and processing of the model.
The weather timing correlation feature extraction module 130 is configured to arrange the plurality of weather data unique heat encoding vectors into a one-dimensional input vector, and then obtain a weather timing feature vector through a weather timing feature extractor based on a one-dimensional convolutional neural network model. It should be understood that, considering that each parameter information in the meteorological data has a dynamic time sequence cooperative variation rule in the time dimension, in order to fully express the time sequence variation characteristics of the meteorological data, the plurality of meteorological data single-heat coding vectors need to be arranged into one-dimensional input vectors and then are encoded in a meteorological time sequence characteristic extractor based on a one-dimensional convolutional neural network model, so that time sequence associated characteristic information of the meteorological data is extracted, and thus, the meteorological time sequence characteristic vectors are obtained.
Optionally, in an embodiment of the present application, the weather timing related feature extraction module is configured to: each layer of the weather time sequence feature extractor based on the one-dimensional convolutional neural network model is used for respectively carrying out the following steps on input data in forward transfer of the layer: performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the weather time sequence feature extractor is the one-dimensional input vector, and the output of the last layer of the weather time sequence feature extractor is the weather time sequence feature vector.
The market trade time sequence correlation feature extraction module 140 is configured to arrange market trade data of agricultural products to be predicted for a plurality of days in the predetermined period of time into a market trade data time sequence input vector according to a time dimension, and then obtain a market trade time sequence feature vector through a market trade data time sequence feature extractor based on a one-dimensional convolutional neural network model.
Optionally, in an embodiment of the present application, the market transaction time sequence correlation feature extraction module is configured to: each layer of the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transfer of the layer: performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; after the input of the first layer of the market transaction data time sequence feature extractor is the market transaction data time sequence input vector, the output of the last layer of the market transaction data time sequence feature extractor is the market transaction time sequence feature vector.
The feature fusion module 150 is configured to fuse the weather timing feature vector and the market transaction timing feature vector to obtain a decoded feature vector using a cascading function. It should be understood that, for the market transaction data of the agricultural product to be predicted, the market transaction data also has time-sequential dynamic change feature information in the time dimension, so in the technical scheme of the application, the market transaction data of the agricultural product to be predicted for a plurality of days in the preset time period is further arranged into a time-sequential input vector of the market transaction data according to the time dimension, and then feature mining is performed in a time-sequential feature extractor of the market transaction data based on a one-dimensional convolutional neural network model, so that time-sequential associated feature information of the market transaction data of the agricultural product to be predicted in the time dimension is extracted, and the time-sequential feature vector of the market transaction is obtained.
Optionally, in an embodiment of the present application, the feature fusion module is configured to: fusing the weather timing feature vector and the market transaction timing feature vector using a cascading function to obtain a decoded feature vector;
wherein the cascading function is:
f(X i ,X j )＝Relu(W f [θ(X i ),φ(X j )])
wherein X is i Representing the weather timing feature vector, θ (X i ) Representing the point convolution of the meteorological time sequence feature vector, X j Representing the time sequence feature vector of the market transaction, phi (X j ) Representing the point convolution of the time sequence feature vector of the market transaction, W f Representing the point convolution of the input, relu being the activation function, []Represents a splicing operation, f (X) i ,X j ) Representing the decoded feature vector.
And the price prediction module 160 is configured to perform decoding regression on the decoding feature vector through a decoder to obtain a decoded value, where the decoded value is used to represent a price prediction value of the agricultural product to be predicted. It should be appreciated that the decoded feature vector is subjected to a decoding regression by a decoder to obtain a decoded value representing the price prediction value of the agricultural product to be predicted. That is, the market transaction data of the agricultural product and the time sequence of the weather data are cooperatively associated with the change characteristic information to decode, thereby predicting the price of the agricultural product.
The price pre-warning module 170 is configured to determine whether to generate price pre-warning prompt information based on a comparison between the decoded value and a predetermined threshold.
Optionally, in an embodiment of the present application, price pre-warning prompt information is generated in response to the decoded value being greater than the predetermined threshold.
Fig. 2 is a schematic structural diagram of a training module in the agricultural product market price early warning management system based on big data analysis according to the embodiment of the application. As shown in fig. 2, in an embodiment of the present application, the agricultural product market price pre-warning management system 100 based on big data analysis further includes a training module 200 for training the weather timing feature extractor based on the one-dimensional convolutional neural network model, the market transaction data timing feature extractor based on the one-dimensional convolutional neural network model, and the decoder, and the training module 200 includes: the training data acquisition module 210 is configured to acquire training data, where the training data includes training meteorological data and training market transaction data of agricultural products to be predicted for a plurality of days within a predetermined period of time, and a true value of a price prediction value of the agricultural products to be predicted; the training independent heat encoding module 220 is configured to independently encode the training meteorological data of each day to obtain a plurality of independent heat encoding vectors of the training meteorological data; the training weather time sequence associated feature extraction module 230 is configured to arrange the plurality of training weather data single-heat encoding vectors into a training one-dimensional input vector, and then obtain a training weather time sequence feature vector through the weather time sequence feature extractor based on the one-dimensional convolutional neural network model; the training market trade time sequence correlation feature extraction module 240 is configured to arrange training market trade data of agricultural products to be predicted for a plurality of days in the predetermined period of time into training market trade data time sequence input vectors according to a time dimension, and then obtain training market trade time sequence feature vectors through the market trade data time sequence feature extractor based on the one-dimensional convolutional neural network model; a training feature fusion module 250 for fusing the training weather timing feature vector and the training market transaction timing feature vector using a cascading function to obtain a training decoding feature vector; a decoding loss module 260, configured to pass the training decoding feature vector through the decoder to obtain a decoding loss function value; and a model training module 270 for training the weather timing feature extractor based on the one-dimensional convolutional neural network model, the market transaction data timing feature extractor based on the one-dimensional convolutional neural network model, and the decoder based on the decoding loss function value and traveling in the direction of gradient descent, wherein in each round of iteration of the training, a weight matrix of the decoder is subjected to half-space structured constraint iteration of weight eigensupport.
Optionally, in an embodiment of the present application, the decoding loss module includes: the training decoding unit is used for carrying out decoding regression on the training decoding characteristic vector by using the decoder according to the following formula so as to obtain a training decoding value; wherein the formula isWherein X is the training decoding eigenvector, Y is the training decoding value, W is a weight matrix,>and a loss function calculation unit for calculating a mean square error between the training decoded value and a true value of the price prediction value of the agricultural product to be predicted to obtain a decoding loss function value.
In particular, in the technical solution of the present application, considering that the weather timing feature vector and the market transaction timing feature vector respectively express the independent thermal coding of the weather data and the time sequence local correlation feature of the market transaction data, when the weather timing feature vector and the market transaction timing feature vector are fused by using a cascade function, the fitting difference, such as over fitting or under fitting, of the weather timing feature vector and the market transaction timing feature vector respectively, may be caused by the point convolution operation, so that the feature distribution of each of the weather timing feature vector and the market transaction timing feature vector to be subjected to the cascade operation has different weight fitting directions with respect to the corresponding part of the weight matrix of the decoder, and thus, the overall feature distribution of the decoded feature vector obtained after the cascade has a problem of poor convergence with respect to the weight matrix of the decoder, thereby affecting the training speed of the decoder. Based on this, the applicant of the present application performs a half-space structuring constraint of the weight matrix M for weight eigensupport during the training process of the decoding eigenvectors, for example denoted V, during each iteration of the weight matrix of the decoder, for example denoted M.
Optionally, in an embodiment of the present application, in each iteration of the training, performing a half-space structured constraint iteration of weight eigen support on a weight matrix of the decoder according to the following optimization formula; wherein, the optimization formula is:
wherein M is the weight matrix of the decoder, V e Is a matrix M T An eigenvector set consisting of eigenvalues of M,and->Respectively representing matrix multiplication and addition, M' represents the weight matrix of the decoder after iteration.
Here, the weight eigen-supported half-space structuring constraint supports correlation integration of eigenvalue sets of a structuring matrix of a weight matrix M of the decoder and a decoding eigenvector V to be decoded, and performs structural support constraint of a hyperplane as a decision boundary on a half-space (half-space) for coupling with a high-dimensional manifold of the decoding eigenvector V represented by the weight matrix M, so that the high-dimensional manifold of the decoding eigenvector V to be decoded can be effectively converged with respect to the hyperplane in a half-space open domain represented by the weight matrix M, thereby improving training speed of the decoder. Therefore, the price prediction of the agricultural products can be accurately performed, and the price early warning is performed on the agricultural products with excessive price, so that the market price stability of the agricultural products is ensured, and powerful support is provided for the macroscopic regulation and control of the government and the appointed production strategy of farmers.
In summary, the agricultural product market price early warning management system based on big data analysis provided by the application constructs an agricultural product market price prediction model based on market transaction data and meteorological data of agricultural products to be predicted so as to generate agricultural product market price early warning information, specifically, the agricultural product market price early warning management system obtains full expression of time sequence cooperative correlation characteristic information between time sequence dynamic change characteristics of the meteorological data and time sequence dynamic change characteristics of the market transaction data of the agricultural products to be predicted through an artificial intelligence technology, so that price prediction of the agricultural products is accurately carried out, price early warning is carried out on the agricultural products with excessive price, market price stability of the agricultural products is further ensured, and powerful support is provided for government macro regulation and farmer specified production strategies.
Fig. 3 is a schematic flow chart of an agricultural product market price early warning management method based on big data analysis according to an embodiment of the present application. As shown in fig. 3, the method includes: s110, meteorological data and market transaction data of agricultural products to be predicted are obtained for a plurality of days in a preset time period; s120, performing independent heat coding on the meteorological data of each day to obtain a plurality of meteorological data independent heat coding vectors; s130, arranging the plurality of meteorological data single-heat coding vectors into one-dimensional input vectors, and then obtaining meteorological time sequence feature vectors through a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model; s140, after market transaction data of agricultural products to be predicted for a plurality of days in the preset time period are arranged into a time sequence input vector of the market transaction data according to the time dimension, the time sequence input vector of the market transaction data is obtained through a time sequence feature extractor of the market transaction data based on a one-dimensional convolutional neural network model; s150, fusing the meteorological time sequence feature vector and the market transaction time sequence feature vector by using a cascading function to obtain a decoding feature vector; s160, carrying out decoding regression on the decoding feature vector through a decoder to obtain a decoding value, wherein the decoding value is used for representing a price prediction value of the agricultural product to be predicted; and S170, determining whether to generate price early warning prompt information based on the comparison between the decoding value and a preset threshold value.
Fig. 4 is a schematic diagram of a model architecture of an agricultural product market price early warning management method based on big data analysis according to an embodiment of the present application. As shown in fig. 4, the model architecture is input with meteorological data and market transaction data of agricultural products to be predicted for a plurality of days within a predetermined period of time. Firstly, performing single-heat coding on the meteorological data of each day to obtain a plurality of single-heat coding vectors of the meteorological data, arranging the single-heat coding vectors of the meteorological data into one-dimensional input vectors, and then obtaining meteorological time sequence feature vectors through a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model. And meanwhile, after the market transaction data of the agricultural products to be predicted for a plurality of days in the preset time period are arranged into the time sequence input vector of the market transaction data according to the time dimension, the time sequence feature vector of the market transaction data is obtained through the time sequence feature extractor of the market transaction data based on the one-dimensional convolutional neural network model. The weather timing feature vector and the market transaction timing feature vector are then fused using a cascading function to obtain a decoded feature vector. And then, carrying out decoding regression on the decoding eigenvector through a decoder to obtain a decoding value, wherein the decoding value is used for representing the price prediction value of the agricultural product to be predicted. And finally, determining whether to generate price early warning prompt information or not based on the comparison between the decoding value and a preset threshold value.
Optionally, in an embodiment of the present application, the arrangement of the plurality of weather data independent heat encoding vectors into a one-dimensional input vector is followed by a weather timing feature extractor based on a one-dimensional convolutional neural network model to obtain a weather timing feature vector, including: each layer of the weather time sequence feature extractor based on the one-dimensional convolutional neural network model is used for respectively carrying out the following steps on input data in forward transfer of the layer: performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the weather time sequence feature extractor is the one-dimensional input vector, and the output of the last layer of the weather time sequence feature extractor is the weather time sequence feature vector.
Optionally, in an embodiment of the present application, the market transaction time sequence correlation feature extraction module is configured to: each layer of the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transfer of the layer: performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; after the input of the first layer of the market transaction data time sequence feature extractor is the market transaction data time sequence input vector, the output of the last layer of the market transaction data time sequence feature extractor is the market transaction time sequence feature vector.
Optionally, in an embodiment of the present application, after the market transaction data of the agricultural products to be predicted for a plurality of days in the predetermined period is arranged into the time dimension of the time series input vector of the market transaction data, the time series feature extractor of the market transaction data based on the one-dimensional convolutional neural network model is used to obtain the time series feature vector of the market transaction, which includes: fusing the weather timing feature vector and the market transaction timing feature vector using a cascading function to obtain a decoded feature vector; wherein the cascading function is:
f(X i ,X j )＝Relu(W f [θ(X i ),φ(X j )])
wherein X is i Representing the weather timing feature vector, θ (X i ) Representing the point convolution of the meteorological time sequence feature vector, X j Representing the time sequence feature vector of the market transaction, phi (X j ) Representing the point convolution of the time sequence feature vector of the market transaction, W f Representing the point convolution of the input, relu being the activation function, []Represents a splicing operation, f (X) i ,X j ) Representing the decoded feature vector.
Fig. 5 is a schematic flow chart of a training phase in an agricultural product market price early warning management method based on big data analysis according to an embodiment of the present application. As shown in fig. 5, the agricultural product market price early warning management method based on big data analysis further includes a training stage for training the weather timing feature extractor based on the one-dimensional convolutional neural network model, the market transaction data timing feature extractor based on the one-dimensional convolutional neural network model, and the decoder. The training phase comprises: s210, a training data acquisition module, which is used for acquiring training data, wherein the training data comprise training meteorological data for a plurality of days in a preset time period and training market transaction data of agricultural products to be predicted, and a true value of a price prediction value of the agricultural products to be predicted; s220, a training independent-heat coding module is used for independent-heat coding the training meteorological data of each day to obtain a plurality of training meteorological data independent-heat coding vectors; s230, a training weather time sequence associated feature extraction module is used for arranging the plurality of training weather data single-heat coding vectors into a training one-dimensional input vector and then obtaining a training weather time sequence feature vector through a weather time sequence feature extractor based on the one-dimensional convolutional neural network model; s240, training market transaction time sequence associated feature extraction module, which is used for arranging training market transaction data of agricultural products to be predicted for a plurality of days in the preset time period into training market transaction data time sequence input vectors according to time dimension, and then obtaining training market transaction time sequence feature vectors through the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model; s250, a training feature fusion module, which is used for fusing the training meteorological time sequence feature vector and the training market transaction time sequence feature vector by using a cascading function to obtain a training decoding feature vector; s260, a decoding loss module, which is used for passing the training decoding characteristic vector through the decoder to obtain a decoding loss function value; and S270, a model training module is used for training the weather time sequence feature extractor based on the one-dimensional convolutional neural network model, the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model and the decoder based on the decoding loss function value and through gradient descent direction propagation, wherein in each round of iteration of training, a weight matrix of the decoder is subjected to half-space structuring constraint iteration of weight intrinsic support.
Fig. 6 is a schematic diagram of a model architecture of a training stage in the agricultural product market price early warning management method based on big data analysis according to the embodiment of the application. As shown in fig. 6, the model architecture of the training phase is input with training meteorological data and training market transaction data of agricultural products to be predicted for a plurality of days within a predetermined period of time, and a true value of the price prediction value of the agricultural products to be predicted. Firstly, performing independent heat coding on the training meteorological data of each day to obtain a plurality of training meteorological data independent heat coding vectors, arranging the plurality of training meteorological data independent heat coding vectors into a training one-dimensional input vector, and then obtaining a training meteorological time sequence feature vector through a meteorological time sequence feature extractor based on the one-dimensional convolutional neural network model. And meanwhile, training market transaction data of the agricultural products to be predicted for a plurality of days in the preset time period are arranged into training market transaction data time sequence input vectors according to the time dimension, and then the training market transaction time sequence feature vectors are obtained through the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model. The training meteorological time series feature vector and the training market transaction time series feature vector are then fused using a cascading function to obtain a training decoding feature vector. The training decoded feature vectors are then passed through the decoder to obtain decoding loss function values. Finally, training the weather timing characteristic extractor based on the one-dimensional convolutional neural network model, the market transaction data timing characteristic extractor based on the one-dimensional convolutional neural network model and the decoder based on the decoding loss function value and through gradient descent direction propagation, wherein in each round of iteration of training, a weight matrix of the decoder is subjected to half-space structuring constraint iteration of weight intrinsic support.
Optionally, in an embodiment of the present application, the decoding loss module includes: performing decoding regression on the training decoding eigenvector by using the decoder in the following formula to obtain a training decoding value;
wherein the formula isWherein X is the training decoding eigenvector, Y is the training decoding value, W is a weight matrix,>representing a matrix multiplication and calculating a mean square error between the training decoded value and a true value of the price prediction value of the agricultural product to be predicted to obtain a decoding loss function value.
Optionally, in an embodiment of the present application, in each iteration of the training, performing a half-space structured constraint iteration of weight eigen support on a weight matrix of the decoder according to the following optimization formula;
wherein, the optimization formula is:
wherein M is the weight matrix of the decoder, V e Is a matrix M T An eigenvector set consisting of eigenvalues of M,and->Respectively representing matrix multiplication and addition, M' represents the weight matrix of the decoder after iteration.
Optionally, in an embodiment of the present application, price pre-warning prompt information is generated in response to the decoded value being greater than the predetermined threshold.
Here, it will be understood by those skilled in the art that the specific operations of the respective steps in the above-described agricultural product market price pre-warning management method based on big data analysis have been described in detail in the above description of the agricultural product market price pre-warning management system based on big data analysis with reference to fig. 1 to 5, and thus, repetitive descriptions thereof will be omitted.
The embodiment of the application also provides a chip system, which comprises at least one processor, and when the program instructions are executed in the at least one processor, the method provided by the embodiment of the application is realized.
The embodiment of the application also provides a computer storage medium, on which a computer program is stored, which when executed by a computer causes the computer to perform the method of the above-described method embodiment.
The present application also provides a computer program product comprising instructions which, when executed by a computer, cause the computer to perform the method of the method embodiment described above.
In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present application, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a digital video disc (digital video disc, DVD)), or a semiconductor medium (e.g., a Solid State Disk (SSD)), or the like.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.

Claims (10)

1. An agricultural product market price early warning management system based on big data analysis, which is characterized by comprising:
the data acquisition module is used for acquiring meteorological data and market transaction data of agricultural products to be predicted for a plurality of days in a preset time period;
the independent heat coding module is used for respectively carrying out independent heat coding on the meteorological data of each day to obtain a plurality of meteorological data independent heat coding vectors;
the meteorological time sequence associated feature extraction module is used for arranging the plurality of meteorological data single-heat coding vectors into one-dimensional input vectors and then obtaining meteorological time sequence feature vectors through a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model;
The market transaction time sequence correlation feature extraction module is used for arranging market transaction data of agricultural products to be predicted for a plurality of days in the preset time period into a market transaction data time sequence input vector according to a time dimension, and obtaining a market transaction time sequence feature vector through a market transaction data time sequence feature extractor based on a one-dimensional convolutional neural network model;
the feature fusion module is used for fusing the meteorological time sequence feature vector and the market transaction time sequence feature vector by using a cascading function to obtain a decoding feature vector;
the price prediction module is used for carrying out decoding regression on the decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is used for representing a price prediction value of the agricultural product to be predicted; and
and the price early warning module is used for determining whether to generate price early warning prompt information or not based on the comparison between the decoding value and a preset threshold value.
2. The blockchain-based data processing method of claim 1, wherein the weather timing correlation feature extraction module is configured to: each layer of the weather time sequence feature extractor based on the one-dimensional convolutional neural network model is used for respectively carrying out the following steps on input data in forward transfer of the layer:
Performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map;
pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images;
non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map;
the input of the first layer of the weather time sequence feature extractor is the one-dimensional input vector, and the output of the last layer of the weather time sequence feature extractor is the weather time sequence feature vector.
3. The blockchain-based data processing method of claim 2, wherein the market transaction time series associated feature extraction module is configured to: each layer of the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transfer of the layer:
performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map;
pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images;
non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map;
after the input of the first layer of the market transaction data time sequence feature extractor is the market transaction data time sequence input vector, the output of the last layer of the market transaction data time sequence feature extractor is the market transaction time sequence feature vector.
4. The blockchain-based data processing method of claim 3, wherein the feature fusion module is configured to: fusing the weather timing feature vector and the market transaction timing feature vector using a cascading function to obtain a decoded feature vector;
wherein the cascading function is:
f(X i ,X j )＝Relu(W f [θ(X i ),φ(X j )])
wherein X is i Representing the weather timing feature vector, θ (X i ) Representing the point convolution of the meteorological time sequence feature vector, X j Representing the time sequence of the market tradeSyndrome vector, phi (X) j ) Representing the point convolution of the time sequence feature vector of the market transaction, W f Representing the point convolution of the input, relu being the activation function, []Represents a splicing operation, f (X) i ,X j ) Representing the decoded feature vector.
5. The blockchain-based data processing method of claim 4, further comprising a training module for training the one-dimensional convolutional neural network model-based weather timing feature extractor, the one-dimensional convolutional neural network model-based market transaction data timing feature extractor, and the decoder.
6. The blockchain-based data processing method of claim 5, wherein the training module includes:
The system comprises a training data acquisition module, a data processing module and a data processing module, wherein the training data acquisition module is used for acquiring training data, the training data comprises training meteorological data for a plurality of days in a preset time period, training market transaction data of agricultural products to be predicted and a true value of a price prediction value of the agricultural products to be predicted;
the training independent heat coding module is used for independent heat coding the training meteorological data of each day to obtain a plurality of training meteorological data independent heat coding vectors;
the training weather time sequence associated feature extraction module is used for arranging the plurality of training weather data single-heat coding vectors into training one-dimensional input vectors and then obtaining training weather time sequence feature vectors through the weather time sequence feature extractor based on the one-dimensional convolutional neural network model;
the training market transaction time sequence correlation feature extraction module is used for arranging training market transaction data of agricultural products to be predicted for a plurality of days in the preset time period into training market transaction data time sequence input vectors according to time dimension, and then obtaining training market transaction time sequence feature vectors through the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model;
the training feature fusion module is used for fusing the training meteorological time sequence feature vector and the training market transaction time sequence feature vector by using a cascading function to obtain a training decoding feature vector;
A decoding loss module, configured to pass the training decoding feature vector through the decoder to obtain a decoding loss function value;
the model training module is used for training the weather time sequence feature extractor based on the one-dimensional convolutional neural network model, the market transaction data time sequence feature extractor based on the one-dimensional convolutional neural network model and the decoder based on the decoding loss function value and through gradient descent direction propagation, wherein in each round of iteration of training, a weight matrix of the decoder is subjected to half-space structuring constraint iteration of weight intrinsic support.
7. The blockchain-based data processing method of claim 6, wherein the decoding loss module includes:
the training decoding unit is used for carrying out decoding regression on the training decoding characteristic vector by using the decoder according to the following formula so as to obtain a training decoding value;
wherein the formula isWherein X is the training decoding eigenvector, Y is the training decoding value, W is a weight matrix,>representing matrix multiplication, and
and the loss function calculation unit is used for calculating the mean square error between the training decoding value and the true value of the price prediction value of the agricultural product to be predicted so as to obtain a decoding loss function value.
8. The blockchain-based data processing method of claim 7, wherein in each iteration of the training, the weight matrix of the decoder is iterated with a half-space structured constraint that is intrinsically supported by weights in the following optimization formula;
wherein, the optimization formula is:
wherein M is the weight matrix of the decoder, V e Is a matrix M T An eigenvector set consisting of eigenvalues of M,and->Respectively representing matrix multiplication and addition, M' represents the weight matrix of the decoder after iteration.
9. The blockchain-based data processing method of claim 8, wherein price pre-warning hints are generated in response to the decoded value being greater than the predetermined threshold.
10. The utility model provides an agricultural product market price early warning management method based on big data analysis, which is characterized by comprising the following steps:
acquiring meteorological data and market transaction data of agricultural products to be predicted for a plurality of days within a preset time period;
the weather data of each day are subjected to independent heat coding respectively to obtain a plurality of weather data independent heat coding vectors;
the plurality of meteorological data independent-heat coding vectors are arranged into one-dimensional input vectors and then a meteorological time sequence feature extractor based on a one-dimensional convolutional neural network model is used for obtaining meteorological time sequence feature vectors;
The market transaction data of the agricultural products to be predicted for a plurality of days in the preset time period are arranged into a time sequence input vector of the market transaction data according to the time dimension, and then the time sequence input vector of the market transaction data is obtained through a time sequence feature extractor of the market transaction data based on a one-dimensional convolutional neural network model;
fusing the weather timing feature vector and the market transaction timing feature vector using a cascading function to obtain a decoded feature vector;
carrying out decoding regression on the decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is used for representing a price prediction value of the agricultural product to be predicted; and
and determining whether to generate price pre-warning prompt information based on the comparison between the decoding value and a preset threshold value.
CN202310637249.2A 2023-05-31 2023-05-31 Agricultural product market price early warning management system and method based on big data analysis Withdrawn CN116664213A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202310637249.2A CN116664213A (en) 2023-05-31 2023-05-31 Agricultural product market price early warning management system and method based on big data analysis
CN202311247179.6A CN117094782A (en) 2023-05-31 2023-09-25 Agricultural product market price early warning management system based on big data analysis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310637249.2A CN116664213A (en) 2023-05-31 2023-05-31 Agricultural product market price early warning management system and method based on big data analysis

Publications (1)

Publication Number Publication Date
CN116664213A true CN116664213A (en) 2023-08-29

Family

ID=87720205

Family Applications (2)

Application Number Title Priority Date Filing Date
CN202310637249.2A Withdrawn CN116664213A (en) 2023-05-31 2023-05-31 Agricultural product market price early warning management system and method based on big data analysis
CN202311247179.6A Withdrawn CN117094782A (en) 2023-05-31 2023-09-25 Agricultural product market price early warning management system based on big data analysis

Family Applications After (1)

Application Number Title Priority Date Filing Date
CN202311247179.6A Withdrawn CN117094782A (en) 2023-05-31 2023-09-25 Agricultural product market price early warning management system based on big data analysis

Country Status (1)

Country Link
CN (2) CN116664213A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117911123A (en) * 2024-03-20 2024-04-19 华高数字科技有限公司 Agricultural product futures transaction supervision system and method based on Internet of things and big data analysis

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117911123A (en) * 2024-03-20 2024-04-19 华高数字科技有限公司 Agricultural product futures transaction supervision system and method based on Internet of things and big data analysis
CN117911123B (en) * 2024-03-20 2024-06-04 华高数字科技有限公司 Agricultural product futures transaction supervision system and method based on Internet of things and big data analysis

Also Published As

Publication number Publication date
CN117094782A (en) 2023-11-21

Similar Documents

Publication Publication Date Title
CN112784920A (en) Cloud-side-end-coordinated dual-anti-domain self-adaptive fault diagnosis method for rotating part
CN116977001A (en) Geological disaster prevention and treatment engineering cost management system and method thereof
CN111310918B (en) Data processing method, device, computer equipment and storage medium
CN117094782A (en) Agricultural product market price early warning management system based on big data analysis
CN115130894A (en) Production planning method and device based on artificial intelligence, computer equipment and medium
CN112784157A (en) Training method of behavior prediction model, behavior prediction method, device and equipment
CN115689443A (en) Supply chain management method and system based on volatility clustering and block chain
CN113656707B (en) Financial product recommendation method, system, storage medium and equipment
CN113259163B (en) Web service quality prediction method and system based on network topology perception
CN112785479B (en) Image invisible watermark universal detection method based on few sample learning
CN112232576B (en) Decision prediction method, device, electronic equipment and readable storage medium
US20210365470A1 (en) Apparatus for recommending feature and method for recommending feature using the same
CN116740567A (en) Soil moisture content detection method and system for Paulownia seedling cultivation
CN116562357B (en) Click prediction model training method and device
CN116628594A (en) User information management system and method based on AI intelligence
CN117972538A (en) Method, device, processor and storage medium for realizing data asset health evaluation processing based on antagonistic neural network
Chen et al. Dynamic comprehensive quality assessment of post-harvest grape in different transportation chains using SAHP–CatBoost machine learning
CN111737319B (en) User cluster prediction method, device, computer equipment and storage medium
CN115238645A (en) Asset data identification method and device, electronic equipment and computer storage medium
CN114757391A (en) Service quality prediction method based on network data space design
CN113901175A (en) Article relation judging method and device
CN118714028B (en) Wavelet decomposition and improved LSTM-based computational power network flow prediction method
CN117725979B (en) Model training method and device, electronic equipment and computer readable storage medium
CN116737607B (en) Sample data caching method, system, computer device and storage medium
KR102696928B1 (en) A method of testing software for potential errors

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication

Application publication date: 20230829

WW01 Invention patent application withdrawn after publication