CN117670413A - Market crowd behavior-based market prediction method - Google Patents
Market crowd behavior-based market prediction method Download PDFInfo
- Publication number
- CN117670413A CN117670413A CN202311706917.9A CN202311706917A CN117670413A CN 117670413 A CN117670413 A CN 117670413A CN 202311706917 A CN202311706917 A CN 202311706917A CN 117670413 A CN117670413 A CN 117670413A
- Authority
- CN
- China
- Prior art keywords
- market
- crowd
- group
- transaction
- data
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000006399 behavior Effects 0.000 claims abstract description 90
- 230000008859 change Effects 0.000 claims abstract description 15
- 238000004458 analytical method Methods 0.000 claims abstract description 8
- 230000007246 mechanism Effects 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims description 27
- 238000012546 transfer Methods 0.000 claims description 22
- 238000011156 evaluation Methods 0.000 claims description 15
- 238000007619 statistical method Methods 0.000 claims description 8
- 238000012935 Averaging Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 4
- 230000000630 rising effect Effects 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000007480 spreading Effects 0.000 claims description 3
- 230000006735 deficit Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 description 6
- 230000003542 behavioural effect Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000007621 cluster analysis Methods 0.000 description 4
- 238000000611 regression analysis Methods 0.000 description 4
- 238000012731 temporal analysis Methods 0.000 description 4
- 238000000700 time series analysis Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000000205 computational method Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010921 in-depth analysis Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
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
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a market prediction method based on market crowd behaviors, which relates to the technical field of market prediction and comprises the following steps: s1, acquiring historical market data; s2, calculating the fluctuation rate R of the market by carrying out statistics and analysis on historical market data; s3, dividing the crowd in the historical market data according to categories, and analyzing the relevance among the crowd in different categories to obtain the relevance P of the market crowd; s4, calculating the weight W of each group of people according to the participation degree and influence of the people of different groups in the historical market data; s5, calculating a random value re of the simulated market crowd according to the fluctuation rate R of the market, the correlation P of the market crowd and the weight W of each crowd; s6, predicting market behaviors according to the random value re of the simulated market crowd and by combining a trust mechanism of the crowd, and obtaining a prediction result. The invention fully considers the relation among market groups, the difference of environment and the change of demand, and predicts the market behaviors more accurately.
Description
Technical Field
The invention relates to the technical field of market prediction, in particular to a market prediction method based on market crowd behaviors.
Background
Market prediction refers to predicting demand and price changes of commodities by analyzing market data and trends to help enterprises make accurate decisions. It relates to a number of fields including data analysis, machine learning, statistics, economics, etc.
By collecting and collating a large amount of market data, such as sales data, price data, supply chain data, etc., an in-depth analysis of the market can be performed. Data analysis can help identify key factors and trends in the market, providing basis for predictions. Machine learning algorithms can predict future market trends by learning historical data and modeling. Common machine learning algorithms include regression analysis, time series analysis, cluster analysis, and the like. The algorithms can be selected and adjusted according to different requirements and data characteristics. Statistical methods can infer overall characteristics and laws through analysis of sample data. For example, by investigating and sampling a number of consumers, a prediction of the consumption behavior and trends of the entire market can be derived. Statistical methods can help to eliminate noise and contingencies in the data, improving the reliability of the predictions. The economics theory can help understand the influence of factors such as supply and demand relation, price forming mechanism and the like of the market on the commodity market. By applying the theory of economics, market data can be better understood and interpreted, and the accuracy of prediction is improved.
However, existing market prediction techniques tend to focus only on the trends and demands of the overall market, and ignore the differences and relationships between different populations. Different populations have different consumption preferences, purchasing power and patterns of behavior, as well as different demands for goods and price sensitivity. The existing prediction technology often regards the crowd as a whole, and ignores the difference among the crowd.
The invention patent with the Chinese application number of 202310302424.2 discloses a user demand prediction method based on big data, which predicts the future demand of a user by analyzing factors such as user behaviors, market trends, competitors and the like, comprehensively considers sales promotion marketing activities, historical demands and other factors and realizes accurate demand prediction. But the prior art still does not take into account differences and changes in the market crowd environment.
Disclosure of Invention
In view of the above, the invention provides a market prediction method based on market crowd behaviors, which fully considers the relationship among market crowd, the difference of environment and the change of demand, and predicts the market behaviors more accurately so as to assist in formulating a more scientific investment strategy.
The technical purpose of the invention is realized as follows:
the invention provides a market prediction method based on market crowd behaviors, which comprises the following steps:
s1, historical market data is obtained, wherein the historical market data comprises crowd behaviors, investment strategies, price fluctuation and crowd quantity of a historical market;
s2, calculating the fluctuation rate R of the market by carrying out statistics and analysis on historical market data;
s3, dividing the crowd in the historical market data according to categories, and analyzing the relevance among the crowd in different categories to obtain the relevance P of the market crowd;
s4, calculating the weight W of each group of people according to the participation degree and influence of the people of different groups in the historical market data;
s5, calculating a random value re of the simulated market crowd according to the fluctuation rate R of the market, the correlation P of the market crowd and the weight W of each crowd;
s6, predicting market behaviors according to the random value re of the simulated market crowd and by combining a trust mechanism of the crowd, and obtaining a prediction result.
Based on the above technical solution, preferably, in step S2, the method for calculating the fluctuation rate of the market is:
r is the fluctuation rate of the market, S i The market price at the time point of the ith historical market data is E is an expected value, delta T is a time interval, and T is the transaction days of one year.
Based on the above technical solution, preferably, step S3 includes:
s31, according to the crowd classification standard, dividing the market participants into a plurality of groups according to categories;
s32, calculating a correlation coefficient between every two groups, linking the groups according to the correlation coefficient to form a link relation between the groups, wherein each link relation comprises a corresponding correlation coefficient value, and the link relation is used as the correlation of market groups;
the calculation formula of the correlation coefficient is as follows:
p is the correlation coefficient, d is the rank difference between every two groups, and n is the number of market participants in a group.
Based on the above technical solution, preferably, step S4 includes:
s41, acquiring behavior data of each group in the historical market data, wherein the behavior data comprise transaction amount, transaction price, transaction frequency, transaction profit and loss and warehouse holding amount;
s42, for each group, calculating the participation degree of each group according to the transaction amount, the transaction frequency and the holding amount; calculating influence of each group according to transaction price, transaction profit and loss and information transfer, wherein the information transfer is obtained by adding information quantity of information propagation according to market participants in each group;
s43, calculating the weight of each group according to the participation degree and the influence.
On the basis of the above technical solution, preferably, step S42 includes:
for a single group a= { a j -j is the number of market participants in the group;
summing to obtain the transaction amount of the group according to the transaction amount of each market participant, calculating the duty ratio of the transaction amount of the group in the transaction amounts of all groups to obtain the transaction amount proportion B of the group A ；
According to each market participationThe transaction frequency of the group is obtained by summing the transaction frequencies of the users in a specified time period A ；
Summing to obtain the holding capacity of the group according to the holding capacity of each market participant, and calculating the duty ratio of the holding capacity of the group in the holding capacities of all groups to obtain the holding capacity proportion D of the group A ；
The participation degree of the group is calculated, and the calculation formula is as follows:
Z A to the participation degree of group A, B A To the transaction amount ratio of group A, C A For group A transaction frequency, D A For the proportion of the holding capacity of the group A, sigma (B+C+D) represents the sum of the transaction capacity proportion, the transaction frequency and the holding capacity proportion of all groups;
summing up and averaging the trading prices of each market participant to obtain the trading price E of the group A ；
According to the trading surplus and shortage of each market participant, summing and averaging to obtain the trading surplus and shortage F of the group A ；
According to the information quantity of each market participant spreading in the market, summarizing to obtain the information transfer value of the market participant, and summing the information transfer values of all the market participants to obtain the information transfer value G of the group A ；
The influence of the group is calculated as follows:
V A for group A influence, E A For group A trade price, F A For the trade surplus and shortage of group A, G A For the information transfer value of group a, Σ (e+f+g) represents the sum of the trading prices, trading earnings and the information transfer value of all groups.
Based on the above technical solution, preferably, in step S43, the method for calculating the weight of each group is as follows:
W A ＝γ×Z A ×V A
gamma is a dynamic coefficient, and the calculation formula is as follows:
γ＝α t×x +(1-α t )×Y t
wherein alpha is a smoothing factor, and the value range is 0-1; t is the sequence number of the time period, t= [1, n]N is the group number of the historical market data grouped according to time period, x is the corresponding historical market data, Y t Is the average value of the historical market data in the t time period, Y t The calculation formula of (2) is as follows:
Y t ＝(X t +X t-1 +...+X 1 +1)/k
wherein X is t For historical market data in the t time period, k is the size of the time window, i.e., the time span of the historical market data.
Based on the above technical solution, preferably, step S5 includes:
the random value of market crowd is re, and its calculation formula is:
wherein R is the fluctuation rate of the market, W is the weight of the crowd, max is the peak value, minPeak value is the minimum peak value, and Q is the number of the crowd.
Based on the above technical solution, preferably, step S6 includes:
s61, for each group of people, each market participant is taken as a node, the trust relationship among each market participant is analyzed according to the crowd behaviors in the historical market data, the trust relationship is represented by connecting lines among the nodes, the trust degree among the market participants is calculated and taken as the weight of the edge, and thus a trust network is constructed;
s62, constructing a market prediction model based on a trust network, taking random values of the crowd as model input, learning crowd behaviors of the crowd, performing behavior adjustment based on trust relations, and outputting prediction results, wherein the prediction results comprise buying and selling behaviors of the crowd, rising and falling of market price and increase and decrease of transaction amount.
On the basis of the above technical solution, preferably, step S61 includes:
for each group of people, extracting the crowd behaviors in the historical market data of the group of people, wherein the crowd behaviors comprise transactions, evaluations and feedback;
taking each market participant as a node;
the trust degree between every two nodes is calculated according to the data of crowd behaviors, and the trust degree calculation formula is as follows:
T(a,b)＝β 1 ×L(a,b)+β 2 ×H(a,b)
t (a, b) represents the trust degree of node a to node b, L (a, b) is the average number of evaluation values of node a obtained by node b in the transaction, H (a, b) is the number of successful transactions between node a and node b, beta 1 And beta 2 Is a emphasis coefficient;
connecting every two market participants to form an edge, and taking the calculated trust degree as the weight of the edge to form an initial network diagram with the weight;
and setting a trust threshold, screening the trust level based on the trust threshold, and deleting the edges with the trust level lower than the trust threshold to obtain a sparse trust network.
On the basis of the above technical solution, preferably, step S62 includes:
collecting market demand data, wherein the market demand data comprises behavior data, transaction data and transaction habits of market participants;
analyzing the market demand data by using a statistical method to obtain market change trend;
and (3) constructing a market prediction model, inputting market demand data, market change trend and a random value re of a simulated market crowd into the market prediction model, and predicting to obtain market behaviors.
Compared with the prior art, the method has the following beneficial effects:
(1) According to the invention, the random value of the simulated market crowd is calculated by calculating the fluctuation rate of the market, analyzing the correlation of the market crowd and calculating the weight of each crowd, and the market behavior is predicted according to the random value and the trust mechanism of the crowd, so that investors are helped to more accurately predict the future trend and behavior of the market, thereby formulating a more scientific investment strategy and improving the success rate and the yield rate of investment;
(2) According to the invention, the historical market data is counted and analyzed, the fluctuation rate of the market is calculated, and the risk level of the market can be estimated. The high fluctuation rate means that market price fluctuation is large and risk is high; the low fluctuation rate indicates smaller fluctuation of market price and lower risk;
(3) The invention can better understand the behaviors and decision process of market participants by analyzing the relevance among different types of market groups. Different categories of market crowd may have different investment preferences, risk bearing capacity and trade behavior, and by analyzing the correlation between them, the dynamics and trends of market participants may be revealed;
(4) According to the invention, the influence of different types of market participants in the market can be known by calculating the weight of each group of people. The higher weighted population represents greater engagement and impact in the market. This may help investors more accurately assess how much different market participants' behavior and decisions affect market price and trend;
(5) According to the invention, the prediction result of the market behaviors can be obtained by simulating the random value of the market crowd and the trust mechanism of the crowd. These predictions can serve as references to investors 'decisions to help them understand better the market's trends and behaviors; these predictions can serve as references to investors 'decisions to help them understand better the market's trends and behaviors. The prediction results may provide the investors with a basis for assisting in decision making. Investors can determine trends and trends in markets based on the forecasted results to determine when to buy, sell or hold the asset. The outcome prediction may also assist the investor in assessing risk and return for different investment scenarios, choosing the most appropriate investment strategy and portfolio.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a calculation result of random values of a simulated market crowd according to an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will clearly and fully describe the technical aspects of the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
As shown in fig. 1, the present invention provides a market prediction method based on market crowd behaviors, including:
s1, historical market data is obtained, wherein the historical market data comprises crowd behaviors, investment strategies, price fluctuation and crowd quantity of a historical market;
s2, calculating the fluctuation rate R of the market by carrying out statistics and analysis on historical market data;
s3, dividing the crowd in the historical market data according to categories, and analyzing the relevance among the crowd in different categories to obtain the relevance P of the market crowd;
s4, calculating the weight W of each group of people according to the participation degree and influence of the people of different groups in the historical market data;
s5, calculating a random value re of the simulated market crowd according to the fluctuation rate R of the market, the correlation P of the market crowd and the weight W of each crowd;
s6, predicting market behaviors according to the random value re of the simulated market crowd and by combining a trust mechanism of the crowd, and obtaining a prediction result.
Specifically, in an embodiment of the present invention, the market may be designated as a commodity market, and step S1 includes:
historical market data in a commodity market is acquired, wherein crowd behavior refers to behavior features and behavior patterns exhibited by market participants in the commodity transaction process. Market participants include producers, consumers, distributors, etc., whose behavior has a significant impact on the supply and demand relationships and price fluctuations of the commodity market. By acquiring crowd behavior data of the historical market, buying will, consumption habits and price sensitivity of market participants can be known, so that trends and trends of the commodity market can be better understood.
Further, the crowd behavior data of the historical market includes:
1. transaction amount and amount: the trading volume and the trading volume are important indicators of trading by market participants, and reflect the activity degree and the fund flow condition of the market. By analyzing the data of the historical transaction amount and the transaction amount, the transaction behaviors and the fund flow directions of market participants can be known, so that the hot spots and the trend of the market can be judged. For example, as the volume and volume of transactions continue to increase, the market may rise; as the volume and amount of transactions continue to decrease, the market may show a tendency to drop.
2. Investors hold warehouse data: the investor holding data can reflect investment preferences and risk preferences of market participants, including multiple void fraction, holding fraction, etc. By analyzing historical investor holding data, the investment direction and risk preference of market participants can be known, and thus market trends and trends can be judged. For example, as the proportion of multiple bins increases, the market may trend upward; as the proportion of empty cartridges increases, the market may show a tendency to drop.
3. Trading behavior of market participants: the trading activity of the market participant includes purchasing, selling, inventory adjustment, and the like. By analyzing the transaction behavior data of the historical market participants, the purchasing strategy and the selling strategy of the market participants can be known, so that the trends and trends of the commodity market can be judged. For example, when market participants generally take purchasing strategies, the market for goods may have a tendency to increase in price; when market participants generally adopt sales strategies, the commodity market may have a tendency to drop in price.
Investment strategies for historic markets: this refers to the investment strategies and operating habits that market participants take in the trade of goods. The investment policy data may include the proportion and frequency of operations to buy, sell, hold, etc.
Price fluctuations in the historic market: this refers to the fluctuation of the market price of the commodity over a period of time. Price volatility data may include the magnitude of the price's rise and fall, volatility, trend of the price, and so forth.
Population of historic market: this refers to the number of people of various types that are involved in the trade with the commodity market for a period of time.
Specifically, in an embodiment of the present invention, step S2 includes:
carrying out statistics and analysis on data such as transaction behaviors, price fluctuation and the like of market participants in the historical market data, and calculating the fluctuation rate R of the market:
r is the fluctuation rate of the market, S i The market price at the time point of the ith historical market data is E is an expected value, delta T is a time interval, and T is the transaction days of one year.
Specifically, in an embodiment of the present invention, step S3 includes:
s31, according to the crowd classification standard, dividing the market participants into a plurality of groups according to categories;
s32, calculating a correlation coefficient between every two groups, linking the groups according to the correlation coefficient to form a link relation between the groups, wherein each link relation comprises a corresponding correlation coefficient value, and the link relation is used as the correlation of market groups;
the calculation formula of the correlation coefficient is as follows:
p is the correlation coefficient, d is the rank difference between every two groups, and n is the number of market participants in a group.
Specifically, in this embodiment, the crowd classification criteria may be classified according to factors such as investment experience, risk preference, and fund size. Each group divided out represents market participants with similar characteristics and behavioral patterns.
After the multiple groups are obtained, the correlation coefficient between each two groups can be calculated to measure the degree of correlation between them. The correlation coefficient is a statistical index used to measure the strength and direction of the linear relationship between two variables. The correlation coefficient has a value ranging from-1 to 1, wherein-1 represents a complete negative correlation, 1 represents a complete positive correlation, and 0 represents no correlation.
In calculating the correlation coefficient, it is first necessary to calculate the rank difference between every two groups. Rank refers to the ranking of market participants in each group by some index, which is then translated into a rank number. Rank difference refers to the difference in rank of the corresponding market participants in the two groups. Then, the squares of all rank differences are added and multiplied by 6, and divided by n multiplied by (n 2-1), the value of the correlation coefficient P is obtained.
The link relationship between groups can be formed by calculating the correlation coefficient between every two groups and taking the correlation number as the weight of the link relationship. Thus, the relevance among different groups can be measured, and the relevance degree among market participants can be known.
Specifically, in an embodiment of the present invention, step S4 includes:
s41, acquiring behavior data of each group in the historical market data, wherein the behavior data comprise transaction amount, transaction price, transaction frequency, transaction profit and loss and warehouse holding amount;
for each group, behavioral data of interest may be extracted from the historical market data. The transaction amount may be obtained by counting the amount of transactions at each time point. The trade price may directly use price information in the historical market data. The transaction frequency may be obtained by counting the number of transactions per group. Trading earnings and earnings can be obtained by calculating the change in the holding and price of each group. The holding amount can be obtained by counting the holding amount at each time point.
S42, for each group, calculating the participation degree of each group according to the transaction amount, the transaction frequency and the holding amount; calculating influence of each group according to transaction price, transaction profit and loss and information transfer, wherein the information transfer is obtained by adding information quantity of information propagation according to market participants in each group;
in this embodiment, step S42 includes:
for a single group a= { a j And j is the number of market participants in the group.
Summing to obtain the transaction amount of the group according to the transaction amount of each market participant, calculating the duty ratio of the transaction amount of the group in the transaction amounts of all groups to obtain the transaction amount proportion B of the group A The method comprises the steps of carrying out a first treatment on the surface of the This ratio may reflect the trading volume duty cycle of the group throughout the market.
Summing up the transaction frequencies C of the group according to the transaction frequency of each market participant in the specified time period A The method comprises the steps of carrying out a first treatment on the surface of the This frequency may reflect the transaction activity level of the group over a specified period of time.
Summing to obtain the holding capacity of the group according to the holding capacity of each market participant, and calculating the duty ratio of the holding capacity of the group in the holding capacities of all groups to obtain the holding capacity proportion D of the group A The method comprises the steps of carrying out a first treatment on the surface of the This ratio may reflect the occupancy duty cycle of the group throughout the market.
The participation degree of the group is calculated, and the calculation formula is as follows:
Z A to the participation degree of group A, B A To the transaction amount ratio of group A, C A For group A transaction frequency, D A For the proportion of the holding capacity of the group A, sigma (B+C+D) represents the sum of the transaction capacity proportion, the transaction frequency and the holding capacity proportion of all groups; the participation degree represents the participation degree of the group in the whole market, and factors such as transaction amount, transaction frequency, warehouse holding amount and the like are comprehensively considered.
Summing up and averaging the trading prices of each market participant to obtain the trading price E of the group A The method comprises the steps of carrying out a first treatment on the surface of the This average price may reflect the transaction price level for the group.
According to the trading surplus and shortage of each market participant, summing and averaging to obtain the trading surplus and shortage F of the group A The method comprises the steps of carrying out a first treatment on the surface of the This average surplus may reflect the profitability of the group.
According to the information quantity of each market participant spreading in the market, summarizing to obtain the information transfer value of the market participant, and summing the information transfer values of all the market participants to obtain the information transfer value G of the group A The method comprises the steps of carrying out a first treatment on the surface of the This value may reflect the information transfer capabilities of the group in the marketplace.
The influence of the group is calculated as follows:
V A for group A influence, E A For group A trade price, F A For the trade surplus and shortage of group A, G A For the information transfer value of group a, Σ (e+f+g) represents the sum of the transaction prices, transaction earnings and earnings, and the information transfer value of all groups. The influence indicates the influence degree of the group in the market, and factors such as transaction price, transaction profit and loss, information transmission and the like are comprehensively considered.
These metrics and computational methods can help to understand the behavior and impact of individual groups in the marketplace. By comparing the participation degree and influence of different groups, main participants and influence in the market can be found, and corresponding investment strategies can be formulated accordingly. In addition, these metrics can also be used to monitor the overall situation and trends of the market, helping investors to better understand market dynamics and make more informed decisions.
S43, calculating the weight of each group according to the participation degree and the influence.
The weight calculation method of each group is as follows:
W A ＝γ×Z A ×V A
gamma is a dynamic coefficient, and the calculation formula is as follows:
γ＝α t×x +(1-α t )×Y t
wherein alpha is a smoothing factor, and the value range is 0-1; t is the sequence number of the time period, t= [1, n]N is the group number of the historical market data grouped according to time period, x is the corresponding historical market data, Y t Is the average value of the historical market data in the t time period, Y t The calculation formula of (2) is as follows:
Y t ＝(X t +X t-1 +...+X 1 +1)/k
wherein X is t For historical market data in the t time period, k is the size of the time window, i.e., the time span of the historical market data.
Calculating the weight of each group according to the participation degree and influence can determine the importance and influence of different groups in the market, so that resources are better distributed and investment strategies are formulated.
The dynamic coefficient gamma is a dynamic adjustment factor calculated according to historical market data and is used for balancing participation degree and influence. In the calculation formula of gamma, alpha is a smoothing factor, and the weights of the groups are calculated by dividing the historical market data into subsets of N groups of historical market data according to a time window k and integrating the degree of time influence. For assessing the importance and impact of different groups in the market. According to the weight, resources and investment emphasis can be more concentrated on the group with higher weight, so as to obtain better investment effect.
By using the calculation method, two factors of participation degree and influence are comprehensively considered, and the actual situation of the group in the market is more accurately reflected by the weight through the adjustment of the dynamic coefficient. This avoids the limitation of evaluating groups based solely on a single indicator of engagement or impact. Meanwhile, through calculation of the average value of historical market data, long-term trend and change of the market can be considered, so that the weight is more stable and reliable.
Specifically, in an embodiment of the present invention, after the volatility R of the market, the correlation P of the market population, and the weight W of each type of population are obtained, the random value re of the simulated market population can be calculated. The calculation formula is as follows:
wherein:
re is a return value representing a market crowd random value (dependent variable).
max is the peak value, representing the maximum value of the market crowd random value.
R is the ripple of the market for controlling the amplitude.
W is the crowd weight for controlling the amplitude.
P is the correlation of market population for controlling wave frequency.
Q is the number of people and represents an argument.
minPeakValue is the minimum peak value used to constrain the minimum value of the market crowd random value.
In this embodiment, the procedure for deriving the formula re is as follows:
firstly, obtaining a sine function of an initial re according to the crowd quantity as an independent variable:
re＝sinQ
where re is in the range of [ -1,1], it is adjusted:
the range is [1, max ].
The embodiment sets P, W, R with respective importance levels, and assigns them with corresponding tasks according to the importance level distinction, where P is used to control the wave frequency, and then converts the re formula into:
while W, R is used to control the amplitude, the re formula continues to be converted into:
to this endControl amplitude, range->P controls the wave frequency, and the range is not discussed.
In this case, the function may have too small an amplitude or even be a straight line, and the peak value is changed because not every peak value of the function is max, and the change has a range, and the range of the return value re is [1, max ], so that in the most extreme case theoretically, the peak value may be 1, and the range of the function is [1,1], that is, a straight line. Based on the above situation, the minimum peak value parameter is set, and then the peak range of the function corresponding to the different parameters is [ minPeakValue,150], and the default value of minPeakValue is 1.
From the above, the amplitude should be controlled to be within the range ofDisassemblingAmplitude control term because of left +.>Is constant, so the right side is adjusted to be within +.>Since the center point of the range of variation is not 0, it can be inferred that there must be a vertical displacement term, considering the following from the standpoint of the range:
sin RW-1
the range is [ -2,0], and the undetermined coefficient M adjusts the range thereof
(sinRW-1)M+2
In this case, the range is [2-2M,2], where M is the range [0,1].
Order the
Is available in the form of
I.e. the amplitude control term is
To sum up:
the market crowd random values of different crowd numbers and crowd characteristics can be obtained by adjusting the values of the parameters and the variables, and the range of the values is [1, max ].
In this embodiment, 100 groups of crowd numbers Q with different numbers are selected as independent variables, random values of the crowd are simulated according to the above formula, and 25 groups with representativeness are selected for display, and specific data are shown in fig. 2.
It should be noted that, in this embodiment, the fluctuation rate R, the crowd weight W, and the correlation P of the market crowd of the market obtained by the calculation are adjusted and constrained, so as to obtain a random value of the market crowd, where the random value re of the market crowd is a result obtained by simulation after fluctuation in a certain range, and the value can represent the behavioral trend of the market crowd to a certain extent, and when the value is used for subsequent prediction, the prediction range can be increased, so as to realize the comprehensiveness of market prediction.
Specifically, in an embodiment of the present invention, step S6 includes:
s61, for each group of people, each market participant is taken as a node, the trust relationship among each market participant is analyzed according to the crowd behaviors in the historical market data, the trust relationship is represented by connecting lines among the nodes, the trust degree among the market participants is calculated and taken as the weight of the edge, and thus a trust network is constructed;
s62, constructing a market prediction model based on a trust network, taking random values of the crowd as model input, learning crowd behaviors of the crowd, performing behavior adjustment based on trust relations, and outputting prediction results, wherein the prediction results comprise buying and selling behaviors of the crowd, rising and falling of market price and increase and decrease of transaction amount.
In this embodiment, step S61 includes:
for each group of people, extracting the crowd behaviors in the historical market data of the group of people, wherein the crowd behaviors comprise transactions, evaluations and feedback; specifically, transaction records, rating scores, rating content, feedback information, and the like may be included.
Taking each market participant as a node; such as buyers, sellers, platforms, etc.
The trust degree between every two nodes is calculated according to the data of crowd behaviors, and the trust degree calculation formula is as follows:
T(a,b)＝β 1 ×L(a,b)+β 2 ×H(a,b)
t (a, b) represents the trust degree of node a to node b, L (a, b) is the average number of evaluation values of node a obtained by node b in the transaction, H (a, b) is the number of successful transactions between node a and node b, beta 1 And beta 2 And the influence of evaluation and transaction times on the trust degree is used as a emphasis coefficient.
In this embodiment, the trust degree of one node to another node is evaluated by comprehensively considering the evaluation and the transaction times by using the trust degree. In the calculation formula of the trust degree, the evaluation value of the node a obtained by the node b in the transaction reflects the performance of the node b in the transaction with the node a. By calculating the average number of evaluation values, the overall performance evaluation of one node b in the transaction can be obtained. The number of successful transactions between node a and node b reflects the history of transactions between node a and node b. By counting the number of successful transactions, the frequency and success rate of transactions between node a and node b can be known. The emphasis coefficient is used for adjusting the influence of evaluation and transaction times on the trust degree, when beta 1 And beta 2 The degree of influence of the evaluation and the number of transactions on the degree of trust may be different when the values of (a) are different. By weighted summing the average number of evaluation values and the number of successful transactions, the trust level (T (a, b)) of node a to node b can be obtained. This confidence value may be used to measure the degree of trust of node a to node b, i.e., the degree of trust of node a to node b.
It should be noted that, the trust level in this embodiment comprehensively considers two factors, namely, the evaluation and the transaction number, so as to evaluate the trust level between the nodes more comprehensively and accurately. By using the formula, a relatively objective and comprehensive trust index can be obtained and used for judging the trust relationship between the nodes.
Connecting every two market participants to form an edge, and taking the calculated trust degree as the weight of the edge to form an initial network diagram with the weight;
and setting a trust threshold, screening the trust level based on the trust threshold, and deleting the edges with the trust level lower than the trust threshold to obtain a sparse trust network.
Specifically, the trust threshold is specifically set according to an application scenario, a system requirement or risk tolerance. For example, if the confidence level is normalized to a value of [0,1], the confidence threshold may be set to 0.4 in consideration of the application scenario and risk tolerance of the commodity market.
By constructing the trust network, the embodiment can exclude market participants with lower trust, and reduce the risk of trading with the market participants. This helps to improve the safety and reliability of the market. Through the trust network, the relationship and the trust degree between market participants can be better understood. This helps to optimize the allocation of market resources, such as more resources to those participants who are more trusted, thereby improving market efficiency and efficiency.
Specifically, in the present embodiment, step S62 includes:
market demand data is collected, including behavioral data, trade habits of market participants.
And analyzing the market demand data by using a statistical method to obtain the market change trend.
And (3) constructing a market prediction model, inputting market demand data, market change trend and a random value re of a simulated market crowd into the market prediction model, and predicting to obtain market behaviors.
The description is given by way of a specific example:
(1) Collecting market demand data
The market demand data may include information such as behavioral data, trade data, and trade habits of market participants. Such data may be obtained by way of market research, consumer surveys, transaction records, and the like.
Market research: market research may employ qualitative and quantitative methods such as face-to-face interviews, questionnaires, focus group discussions, etc. to obtain consumer opinion, feedback, and demand.
Consumer investigation: by issuing a questionnaire to the consumer or conducting an online survey, information such as purchasing habit, purchasing will, consumption budget and the like of the consumer is known.
Transaction records: and acquiring the purchasing behavior data and the transaction data of the market participants by analyzing the transaction records. For example, consumer purchase records may be collected, including information about the products purchased, time of purchase, number of purchases, channel of purchase, and the like.
Market competition analysis: in addition to collecting consumer data, market competitor data may also be collected. The competition status and trend of the market are known by analyzing the market share, product pricing, promotion strategy and other information of the competitors.
(2) Analysis of market demand data using statistical methods
Through statistical analysis of market demand data, the change trend of market demand is known. Statistical methods include descriptive statistics, regression analysis, time series analysis, and the like. By these methods, seasonal changes, long-term trends, etc. of market demand are identified and future market demand is predicted.
Descriptive statistics: descriptive statistics is a method of summarizing and describing market demand data. The average sales level and the sales fluctuation degree of the product are known by calculating the average value and the standard deviation of the sales amount of the product.
Regression analysis: regression analysis is used to study the relationship between market demand data and other variables. By establishing a regression model, the relationship between market demands and factors such as price, sales promotion, competitor sales and the like is known.
Time series analysis: the time series analysis is used for researching the change rule of market demand data along with time. Future market demands are predicted by analyzing characteristics of trends, seasonality, periodicity and the like of the market demand data.
And (3) cluster analysis: cluster analysis is used to divide market demand data into different groups or categories. Through cluster analysis, different market segments or consumer groups present in the market demand data can be found.
(3) Predicting market behavior
The prediction result of the market behavior can be obtained by inputting market demand data, market change trend and random value re of the simulated market crowd into the market prediction model. These forecasts may include information about the market participant's willingness to purchase, purchasing power, number of purchases, etc.
The regression model is adopted as a market prediction model, and the expression form of the regression model is as follows:
L＝δ 0 +δ 1 Γ 1 +δ 2 Γ 2 +...+δ η Γ η +ζ
wherein L is a dependent variable, i.e. market behavior index, Γ 1 ,Γ 2 ,...,Γ η Is an independent variable, namely, market demand data, market change trend and random value of simulated market crowd, delta 0 ,δ 1 ,...,δ η Is the regression coefficient and ζ is the error term.
The market behaviors predicted by the return model at least comprise: 1) Sales over a period of time in the future; 2) Sales volume over a period of time in the future; 3) The rate of increase in market size over a period of time in the future; 4) The popularity of a product or service in a target market for a period of time in the future; 5) Average amount of consumption of market participants over a period of time in the future; 6) Repeated purchasing behavior of market participants over a period of time in the future.
It should be noted that, the random value re of the simulated market crowd is also used as a part of the model input in the embodiment, which can play a role of increasing the prediction range and considering uncertainty, and in particular, the random value of the simulated market crowd can help the model consider the random behaviors and decisions of market participants, so as to more comprehensively predict the market behaviors.
(4) Validating and tuning a model
And verifying and testing the constructed market prediction model by using historical market data, and evaluating the accuracy and stability of the model. If the prediction effect of the model is not ideal, the model can be adjusted and optimized, such as adjusting model parameters, changing model structure, etc., so as to improve the accuracy of prediction.
(5) Monitoring and updating predictions
The prediction accuracy of the market behaviors and the models is continuously monitored, and the market behaviors and the model performances are updated and adjusted according to the market changes and the model performances. Market demand is dynamically changing, so the model needs to be updated in time to accommodate changes in the market. At the same time, there is also a need to monitor changes in the behavior of market participants and changes in the market environment, and to adjust the input and parameters of the model to ensure accuracy of the predictions.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (10)
1. The market prediction method based on the market crowd behaviors is characterized by comprising the following steps of:
s1, historical market data is obtained, wherein the historical market data comprises crowd behaviors, investment strategies, price fluctuation and crowd quantity of a historical market;
s2, calculating the fluctuation rate R of the market by carrying out statistics and analysis on historical market data;
s3, dividing the crowd in the historical market data according to categories, and analyzing the relevance among the crowd in different categories to obtain the relevance P of the market crowd;
s4, calculating the weight W of each group of people according to the participation degree and influence of the people of different groups in the historical market data;
s5, calculating a random value re of the simulated market crowd according to the fluctuation rate R of the market, the correlation P of the market crowd and the weight W of each crowd;
s6, predicting market behaviors according to the random value re of the simulated market crowd and by combining a trust mechanism of the crowd, and obtaining a prediction result.
2. The market prediction method based on market crowd behaviors according to claim 1, wherein in step S2, the calculation method of the fluctuation rate of the market is:
r is the fluctuation rate of the market, S i The market price at the time point of the ith historical market data is E is an expected value, delta T is a time interval, and T is the transaction days of one year.
3. The market prediction method based on market crowd behavior according to claim 1, wherein step S3 includes:
s31, according to the crowd classification standard, dividing the market participants into a plurality of groups according to categories;
s32, calculating a correlation coefficient between every two groups, linking the groups according to the correlation coefficient to form a link relation between the groups, wherein each link relation comprises a corresponding correlation coefficient value, and the link relation is used as the correlation of market groups;
the calculation formula of the correlation coefficient is as follows:
p is the correlation coefficient, d is the rank difference between every two groups, and n is the number of market participants in a group.
4. The market prediction method based on market crowd behavior according to claim 1, wherein step S4 comprises:
s41, acquiring behavior data of each group in the historical market data, wherein the behavior data comprise transaction amount, transaction price, transaction frequency, transaction profit and loss and warehouse holding amount;
s42, for each group, calculating the participation degree of each group according to the transaction amount, the transaction frequency and the holding amount; calculating influence of each group according to transaction price, transaction profit and loss and information transfer, wherein the information transfer is obtained by adding information quantity of information propagation according to market participants in each group;
s43, calculating the weight of each group according to the participation degree and the influence.
5. The market prediction method based on market crowd behavior according to claim 4, wherein step S42 includes:
for a single group a= { a j -j is the number of market participants in the group;
summing up the transaction amounts of each market participant to obtain theCalculating the duty ratio of the transaction amount of the group in the transaction amounts of all groups to obtain the transaction amount proportion B of the group A ；
Summing up the transaction frequencies C of the group according to the transaction frequency of each market participant in the specified time period A ；
Summing to obtain the holding capacity of the group according to the holding capacity of each market participant, and calculating the duty ratio of the holding capacity of the group in the holding capacities of all groups to obtain the holding capacity proportion D of the group A ；
The participation degree of the group is calculated, and the calculation formula is as follows:
Z A to the participation degree of group A, B A To the transaction amount ratio of group A, C A For group A transaction frequency, D A For the proportion of the holding capacity of the group A, sigma (B+C+D) represents the sum of the transaction capacity proportion, the transaction frequency and the holding capacity proportion of all groups;
summing up and averaging the trading prices of each market participant to obtain the trading price E of the group A ；
According to the trading surplus and shortage of each market participant, summing and averaging to obtain the trading surplus and shortage F of the group A ；
According to the information quantity of each market participant spreading in the market, summarizing to obtain the information transfer value of the market participant, and summing the information transfer values of all the market participants to obtain the information transfer value G of the group A ；
The influence of the group is calculated as follows:
V A for group A influence, E A For group A trade price, F A Is a groupTrade surplus and deficit of A, G A For the information transfer value of group a, Σ (e+f+g) represents the sum of the transaction prices, transaction earnings and earnings, and the information transfer value of all groups.
6. The market prediction method based on market crowd behavior according to claim 5, wherein in step S43, the calculation method of the weight of each group is:
W A ＝γ×Z A ×V A
gamma is a dynamic coefficient, and the calculation formula is as follows:
γ＝α t×x +(1-α t )×Y t
wherein alpha is a smoothing factor, and the value range is 0-1; t is the sequence number of the time period, t= [1, n]N is the group number of the historical market data grouped according to time period, x is the corresponding historical market data, Y t Is the average value of the historical market data in the t time period, Y t The calculation formula of (2) is as follows:
Y t ＝(X t +X t-1 +...+X 1 +1)/k
wherein X is t For historical market data in the t time period, k is the size of the time window, i.e., the time span of the historical market data.
7. The market prediction method based on market crowd behavior according to claim 1, wherein step S5 includes:
the random value of market crowd is re, and its calculation formula is:
wherein R is the fluctuation rate of the market, W is the weight of the crowd, max is the peak value, minPeak value is the minimum peak value, and Q is the number of the crowd.
8. The market prediction method based on market crowd behavior according to claim 1, wherein step S6 includes:
s61, for each group of people, each market participant is taken as a node, the trust relationship among each market participant is analyzed according to the crowd behaviors in the historical market data, the trust relationship is represented by connecting lines among the nodes, the trust degree among the market participants is calculated and taken as the weight of the edge, and thus a trust network is constructed;
s62, constructing a market prediction model based on a trust network, taking random values of the crowd as model input, learning crowd behaviors of the crowd, performing behavior adjustment based on trust relations, and outputting prediction results, wherein the prediction results comprise buying and selling behaviors of the crowd, rising and falling of market price and increase and decrease of transaction amount.
9. The market prediction method based on market crowd behavior according to claim 8, wherein step S61 includes:
for each group of people, extracting the crowd behaviors in the historical market data of the group of people, wherein the crowd behaviors comprise transactions, evaluations and feedback;
taking each market participant as a node;
the trust degree between every two nodes is calculated according to the data of crowd behaviors, and the trust degree calculation formula is as follows:
T(a,b)＝β 1 ×L(a,b)+β 2 ×H(a,b)
t (a, b) represents the trust degree of node a to node b, L (a, b) is the average number of evaluation values of node a obtained by node b in the transaction, H (a, b) is the number of successful transactions between node a and node b, beta 1 And beta 2 Is a emphasis coefficient;
connecting every two market participants to form an edge, and taking the calculated trust degree as the weight of the edge to form an initial network diagram with the weight;
and setting a trust threshold, screening the trust level based on the trust threshold, and deleting the edges with the trust level lower than the trust threshold to obtain a sparse trust network.
10. The market prediction method based on market crowd behavior according to claim 8, wherein step S62 includes:
collecting market demand data, wherein the market demand data comprises behavior data, transaction data and transaction habits of market participants;
analyzing the market demand data by using a statistical method to obtain market change trend;
and (3) constructing a market prediction model, inputting market demand data, market change trend and a random value re of a simulated market crowd into the market prediction model, and predicting to obtain market behaviors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311706917.9A CN117670413A (en) | 2023-12-13 | 2023-12-13 | Market crowd behavior-based market prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311706917.9A CN117670413A (en) | 2023-12-13 | 2023-12-13 | Market crowd behavior-based market prediction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117670413A true CN117670413A (en) | 2024-03-08 |
Family
ID=90065975
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311706917.9A Pending CN117670413A (en) | 2023-12-13 | 2023-12-13 | Market crowd behavior-based market prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117670413A (en) |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6792399B1 (en) * | 1999-09-08 | 2004-09-14 | C4Cast.Com, Inc. | Combination forecasting using clusterization |
| US7072863B1 (en) * | 1999-09-08 | 2006-07-04 | C4Cast.Com, Inc. | Forecasting using interpolation modeling |
| US8301548B1 (en) * | 2010-04-09 | 2012-10-30 | Alpha Vision Services, Llc | Methods and systems related to securities trading |
| US8935198B1 (en) * | 1999-09-08 | 2015-01-13 | C4Cast.Com, Inc. | Analysis and prediction of data using clusterization |
| CN109449899A (en) * | 2019-01-14 | 2019-03-08 | 华北电力大学 | A kind of longitudinal protection method based on Spearman rank correlation coefficient |
| CN112083270A (en) * | 2020-08-14 | 2020-12-15 | 昆明理工大学 | Wind power plant current collection line single-phase earth fault line selection method based on correlation coefficient |
| CN112465566A (en) * | 2020-12-14 | 2021-03-09 | 树蛙信息科技(南京)有限公司 | Shopping mall passenger flow volume prediction method based on historical data auxiliary scene analysis |
| CN112765442A (en) * | 2018-06-25 | 2021-05-07 | 中译语通科技股份有限公司 | Network emotion fluctuation index monitoring and analyzing method and system based on news big data |
| CN112884583A (en) * | 2021-03-22 | 2021-06-01 | 杭州蓝目数字信息技术有限公司 | Method for stock trading strategy bionic and evolutionary algorithm |
| CN114117355A (en) * | 2022-01-27 | 2022-03-01 | 平安科技（深圳）有限公司 | Optimization method, system, equipment and readable storage medium of time-varying-resistance model |
| CN114219169A (en) * | 2021-12-23 | 2022-03-22 | 上海颖幡技术有限公司 | Script banner supply chain sales and inventory prediction algorithm model and application system |
| CN116090624A (en) * | 2022-12-30 | 2023-05-09 | 河海大学 | Fine granularity load segmentation prediction method |
| CN116228280A (en) * | 2023-03-27 | 2023-06-06 | 杨彬 | User demand prediction method based on big data |
| CN116886998A (en) * | 2023-07-19 | 2023-10-13 | 中教畅享(北京)科技有限公司 | Interactive processing method for simulating live broadcast environment |
-
2023
- 2023-12-13 CN CN202311706917.9A patent/CN117670413A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7072863B1 (en) * | 1999-09-08 | 2006-07-04 | C4Cast.Com, Inc. | Forecasting using interpolation modeling |
| US8935198B1 (en) * | 1999-09-08 | 2015-01-13 | C4Cast.Com, Inc. | Analysis and prediction of data using clusterization |
| US6792399B1 (en) * | 1999-09-08 | 2004-09-14 | C4Cast.Com, Inc. | Combination forecasting using clusterization |
| US8301548B1 (en) * | 2010-04-09 | 2012-10-30 | Alpha Vision Services, Llc | Methods and systems related to securities trading |
| CN112765442A (en) * | 2018-06-25 | 2021-05-07 | 中译语通科技股份有限公司 | Network emotion fluctuation index monitoring and analyzing method and system based on news big data |
| CN109449899A (en) * | 2019-01-14 | 2019-03-08 | 华北电力大学 | A kind of longitudinal protection method based on Spearman rank correlation coefficient |
| CN112083270A (en) * | 2020-08-14 | 2020-12-15 | 昆明理工大学 | Wind power plant current collection line single-phase earth fault line selection method based on correlation coefficient |
| CN112465566A (en) * | 2020-12-14 | 2021-03-09 | 树蛙信息科技(南京)有限公司 | Shopping mall passenger flow volume prediction method based on historical data auxiliary scene analysis |
| CN112884583A (en) * | 2021-03-22 | 2021-06-01 | 杭州蓝目数字信息技术有限公司 | Method for stock trading strategy bionic and evolutionary algorithm |
| CN114219169A (en) * | 2021-12-23 | 2022-03-22 | 上海颖幡技术有限公司 | Script banner supply chain sales and inventory prediction algorithm model and application system |
| CN114117355A (en) * | 2022-01-27 | 2022-03-01 | 平安科技（深圳）有限公司 | Optimization method, system, equipment and readable storage medium of time-varying-resistance model |
| CN116090624A (en) * | 2022-12-30 | 2023-05-09 | 河海大学 | Fine granularity load segmentation prediction method |
| CN116228280A (en) * | 2023-03-27 | 2023-06-06 | 杨彬 | User demand prediction method based on big data |
| CN116886998A (en) * | 2023-07-19 | 2023-10-13 | 中教畅享(北京)科技有限公司 | Interactive processing method for simulating live broadcast environment |
Non-Patent Citations (6)
| Title |
|---|
| 曹迎春;刘善存;邱菀华;: "证券市场日内流动性的综合度量、特征与信息含量", 系统工程, no. 03, 28 March 2007 (2007-03-28) * |
| 王英立;陶帅;候晓晓;齐宏;: "基于MIV分析的GA-BP神经网络光伏短期发电预测", 太阳能学报, no. 08, 25 August 2020 (2020-08-25) * |
| 蒋锋;彭紫君;: "基于混沌PSO优化BP神经网络的碳价预测", 统计与信息论坛, no. 05, 10 May 2018 (2018-05-10) * |
| 郭永济;张谊浩;: "空气质量会影响股票市场吗?", 金融研究, no. 02, 25 February 2016 (2016-02-25) * |
| 顾银宽;: "B-S模型在可转换债券估价中的应用分析", 安徽工业大学学报(社会科学版), no. 02, 15 March 2009 (2009-03-15) * |
| 饶东宁;郭海峰;蒋志华;: "基于并行概率规划的股票指数模拟", 计算机学报, no. 06, 27 November 2018 (2018-11-27) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Venkatesan et al. | Optimal customer relationship management using Bayesian decision theory: An application for customer selection | |
| Matsatsinis et al. | Intelligent support systems for marketing decisions | |
| Miller et al. | Eliciting informative feedback: The peer-prediction method | |
| JP6034890B2 (en) | Specification, estimation, causal driver discovery and market response elasticity or lift coefficient automation | |
| US8825514B2 (en) | System and method for estimating residual lifetime value of a customer base utilizing survival analysis | |
| Ellison et al. | Costs of managerial attention and activity as a source of sticky prices: Structural estimates from an online market | |
| US7664693B1 (en) | Financial methodology for the examination and explanation of spread between analyst targets and market share prices | |
| US11216850B2 (en) | Predictive platform for determining incremental lift | |
| EP2506207A1 (en) | Computer system, computer-implemented method and computer program product for managing operations of a system | |
| Carrillo et al. | Can tightness in the housing market help predict subsequent home price appreciation? Evidence from the United States and the Netherlands | |
| Maaß et al. | Improving short-term demand forecasting for short-lifecycle consumer products with data mining techniques | |
| CZ20013132A3 (en) | Valuation prediction models in situations with missing inputs | |
| JP2005530232A5 (en) | ||
| JP2013519939A (en) | Method and apparatus for predicting repurchase trends | |
| US11188829B2 (en) | Asymmetrical multilateral decision support system | |
| US11934971B2 (en) | Systems and methods for automatically building a machine learning model | |
| Hogenboom et al. | Adaptive tactical pricing in multi‐agent supply chain markets using economic regimes | |
| US8255316B2 (en) | Integrated business decision-making system and method | |
| CN118469609A (en) | Commodity combined sales method and system based on e-commerce platform | |
| Finlay | Towards profitability: A utility approach to the credit scoring problem | |
| CN112767114A (en) | Enterprise diversified decision method and device, electronic equipment and storage medium | |
| Sari | Development of an integrated discounting strategy based on vendors’ expectations using FAHP and fuzzy goal programming | |
| JP6381844B1 (en) | Computer system, method, and program for accumulating assets whose value varies over time | |
| CN117853158A (en) | Enterprise operation data prediction method and device based on dynamic quantity benefit analysis | |
| CN117670413A (en) | Market crowd behavior-based market prediction method |
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 |