CN111835536A - Flow prediction method and device - Google Patents
Flow prediction method and device Download PDFInfo
- Publication number
- CN111835536A CN111835536A CN201910305627.0A CN201910305627A CN111835536A CN 111835536 A CN111835536 A CN 111835536A CN 201910305627 A CN201910305627 A CN 201910305627A CN 111835536 A CN111835536 A CN 111835536A
- Authority
- CN
- China
- Prior art keywords
- predicted
- time
- flow
- prediction
- moment
- 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.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides a traffic prediction method and a traffic prediction device, which can input real traffic at a time before a time to be predicted in a prediction period to which the time to be predicted belongs to a first pre-trained prediction network model to obtain first predicted traffic, input real traffic at a time corresponding to the time to be predicted in the prediction period to which the time to be predicted belongs to a second pre-trained prediction network model to obtain second predicted traffic, and determine the predicted traffic at the time to be predicted based on the first predicted traffic and the second predicted traffic. In the prior art, the first predicted flow is used as the predicted flow at the time to be predicted, the predicted flow at the time to be predicted is determined based on the first predicted flow and the second predicted flow, and the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, so that the accuracy of a flow prediction result can be improved.
Description
Technical Field
The present application relates to the field of big data analysis technologies, and in particular, to a method and an apparatus for predicting traffic.
Background
With the rapid development of computer technology, predicting the traffic of a specified object at a certain future time (which may be referred to as a time to be predicted) is widely applied in different scenes, for example, predicting the traffic of a certain area, and better managing and controlling the traffic condition of the area according to the predicted traffic of the area; the network flow of a certain website is predicted, and the situation of sudden network flow increase of the website can be timely dealt with according to the size of the predicted network flow.
In the prior art, the traffic of the time to be predicted can be predicted according to the real traffic of the time before the time to be predicted in the prediction cycle to which the time to be predicted belongs and a pre-trained prediction network model. For example, with days as a prediction period, the real pedestrian volume at 9 points in 10 days of 2 months, the real pedestrian volume at 10 points and the real pedestrian volume at 11 points may be input to a pre-trained prediction network model to obtain the predicted pedestrian volume at 12 points in 10 days of 2 months, where the training sample set of the prediction network model may include the pedestrian volume data of days before 10 days of 2 months, and the pedestrian volume data of each day may include the real pedestrian volume at 9 points, the real pedestrian volume at 10 points, the real pedestrian volume at 11 points and the real pedestrian volume at 12 points on the day.
Therefore, in the prior art, the flow at the time to be predicted is predicted only according to the real flow at a plurality of times belonging to the same prediction cycle as the time to be predicted, which may result in a low accuracy of the flow prediction result.
Disclosure of Invention
The embodiment of the application aims to provide a flow prediction method and a flow prediction device, which can improve the accuracy of a flow prediction result. The specific technical scheme is as follows:
in a first aspect, to achieve the above object, an embodiment of the present application discloses a traffic prediction method, where the method includes:
acquiring real flow of a moment before the moment to be predicted in a prediction period to which the moment to be predicted belongs, and real flow of a moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
inputting the real traffic of the time before the time to be predicted into a pre-trained first prediction network model in a prediction cycle to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction cycle according to the real traffic of the times in the same prediction cycle;
inputting the real traffic of the moment corresponding to the moment to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to obtain second prediction traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
Optionally, the determining the predicted flow at the time to be predicted based on the first predicted flow and the second predicted flow includes:
inputting the first predicted flow and the second predicted flow to a pre-trained regression model, wherein the regression model is obtained by training by taking the predicted flows of a plurality of historical moments output by the first predicted network model and the predicted flows of the plurality of historical moments output by the second predicted network model as input data and taking the real flows of the plurality of historical moments as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
Optionally, the determining the predicted flow at the time to be predicted based on the first predicted flow and the second predicted flow includes:
inputting the first predicted flow, the second predicted flow and the time information of the time to be predicted into a regression model trained in advance, wherein the regression model is obtained by training by taking the predicted flows of a plurality of historical moments output by the first predicted network model, the predicted flows of the plurality of historical moments output by the second predicted network model and the time information of the plurality of historical moments as input data and the real flows of the plurality of historical moments as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
Optionally, in the prediction period to which the time to be predicted belongs, before the real flow at the time before the time to be predicted, and before the real flow at the time before the prediction period to which the time to be predicted belongs and before the real flow at the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs, the method further includes:
judging whether the distance between the moment to be predicted and the current moment is greater than a preset duration or not;
and if the distance between the moment to be predicted and the current moment is not more than the preset time length, executing the step of acquiring the real flow of the moment to be predicted in the prediction period before the moment to be predicted and the real flow of the moment corresponding to the moment to be predicted in the prediction period before the moment to be predicted.
Optionally, the method further includes:
if the distance between the moment to be predicted and the current moment is greater than the preset duration, acquiring the real flow of the moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
and inputting the real flow of the time corresponding to the time to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
In a second aspect, in order to achieve the above object, an embodiment of the present application discloses a flow rate prediction apparatus, including:
the acquisition module is used for acquiring the real flow of the time before the time to be predicted in the prediction period to which the time to be predicted belongs and the real flow of the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs;
the first prediction module is used for inputting the real traffic of the time before the time to be predicted into a first pre-trained prediction network model in a prediction period to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction period according to the real traffic of the times in the same prediction period;
the second prediction module is used for inputting the real traffic of the moment corresponding to the moment to be predicted into a second pre-trained prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to so as to obtain second predicted traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and the determining module is used for determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
Optionally, the determining module is specifically configured to input the first predicted traffic and the second predicted traffic to a regression model trained in advance, where the regression model is obtained by training using predicted traffic at a plurality of historical times output by the first predicted network model and predicted traffic at a plurality of historical times output by the second predicted network model as input data, and using real traffic at a plurality of historical times as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
Optionally, the determining module is specifically configured to input the first predicted traffic, the second predicted traffic, and the time information of the time to be predicted into a regression model trained in advance, where the regression model is obtained by training using predicted traffic of a plurality of historical times output by the first predicted network model, predicted traffic of the plurality of historical times output by the second predicted network model, and the time information of the plurality of historical times as input data, and using real traffic of the plurality of historical times as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
Optionally, the apparatus further comprises:
the judging module is used for judging whether the distance between the moment to be predicted and the current moment is greater than a preset duration;
and if the distance between the moment to be predicted and the current moment is not more than the preset time length, triggering the acquisition module.
Optionally, the apparatus further comprises:
the processing module is used for acquiring the real flow of the time corresponding to the time to be predicted in a prediction period before the prediction period to which the time to be predicted belongs if the distance between the time to be predicted and the current time is greater than the preset time length;
and inputting the real flow of the time corresponding to the time to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
In another aspect of this application, in order to achieve the above object, an embodiment of this application further discloses an electronic device, where the electronic device includes a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;
the memory is used for storing a computer program;
the processor is configured to implement the flow prediction method according to the first aspect when executing the program stored in the memory.
In yet another aspect of this embodiment, there is also provided a computer-readable storage medium having stored therein instructions which, when run on a computer, implement the flow prediction method according to the first aspect described above.
In yet another aspect of this embodiment, a computer program product containing instructions is provided, which when executed on a computer, causes the computer to perform the method for flow prediction according to the first aspect.
The embodiment of the application provides a traffic prediction method, which can input real traffic at a time before a time to be predicted in a prediction period to which the time to be predicted belongs to a pre-trained first prediction network model to obtain first predicted traffic, input real traffic at a time corresponding to the time to be predicted in the prediction period to which the time to be predicted belongs to a pre-trained second prediction network model to obtain second predicted traffic, and determine the predicted traffic at the time to be predicted based on the first predicted traffic and the second predicted traffic. In the prior art, the first predicted flow is used as the predicted flow at the time to be predicted, and the application determines the predicted flow at the time to be predicted based on the first predicted flow and the second predicted flow, that is, the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, so that the accuracy of the flow prediction result can be improved based on the method of the application.
Of course, not all advantages described above need to be achieved at the same time in the practice of any one product or method of the present application.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a flow prediction method according to an embodiment of the present application;
fig. 2 is a flowchart of an example of a traffic prediction method according to an embodiment of the present application;
fig. 3 is a structural diagram of a flow rate prediction apparatus according to an embodiment of the present application;
fig. 4 is a structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In the prior art, the flow of the time to be predicted is predicted only according to the real flow of a plurality of times which belong to the same prediction period with the time to be predicted, which may result in low accuracy of the flow prediction result.
In order to solve the above problem, an embodiment of the present application provides a traffic prediction method, which may be applied to an electronic device, where the electronic device may be a terminal or a server, and the electronic device is configured to predict traffic of a specified object. For example, the traffic of a person in a certain area is predicted, or the traffic of a network in a certain site is predicted.
The electronic device may obtain a real traffic (may be referred to as a horizontal historical real traffic) at a time (may be referred to as a horizontal historical time) before a time to be predicted in a prediction cycle to which the time to be predicted belongs and a real traffic (may be referred to as a vertical historical real traffic) at a time (may be referred to as a vertical historical time) corresponding to the time to be predicted in a prediction cycle before the time to be predicted belongs, then, the electronic device may input the horizontal historical real traffic to a first prediction network model trained in advance to obtain a first prediction traffic, input the vertical historical real traffic to a second prediction network model trained in advance to obtain a second prediction traffic, and further, the electronic device may determine the prediction traffic at the time to be predicted based on the first prediction traffic and the second prediction traffic.
According to the method, the flow at the moment to be predicted is predicted through the transverse historical real flow and the longitudinal historical real flow respectively to obtain the first predicted flow and the second predicted flow, the predicted flow at the moment to be predicted is further determined based on the first predicted flow and the second predicted flow, and the influence of the real flow at the moment corresponding to the moment to be predicted in different prediction periods on the flow at the moment to be predicted is comprehensively considered, so that the accuracy of a flow prediction result can be improved based on the method.
Referring to fig. 1, fig. 1 is a flowchart of a flow prediction method provided in an embodiment of the present application, where the method may include the following steps:
s101: and acquiring the real flow of the time before the time to be predicted in the prediction period to which the time to be predicted belongs, and the real flow of the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs.
The duration of the prediction period may be set by a technician according to experience, for example, the day may be used as the prediction period, or the month may be used as the prediction period, but is not limited thereto.
In the actual operation process, the statistical flow rate is usually the total flow rate in a certain time period, and therefore, the time to be predicted may be a time period in the prediction period to which the time to be predicted belongs, for example, if the day is taken as the prediction period, the time to be predicted may be one hour of the day, or half an hour of the day; if a month is taken as the prediction period, the time to be predicted may be one day of a month, or half a day of a month.
In the application implementation, after the electronic device determines the time to be predicted, the electronic device may obtain the horizontal historical real flow of the time to be predicted and the vertical historical real flow of the time to be predicted, and then the electronic device may predict the flow of the time to be predicted from different dimensions according to the horizontal historical real flow and the vertical historical real flow.
In one implementation, if a day is used as a prediction period, and the time to be predicted is 2 months, 10 days, 12 points and 30 points, the electronic device may obtain the real flow of 2 months, 10 days, 10 points and 30 points, the real flow of 10 points and 30 points and 11 points and 30 points and 12 points, as the transverse historical real flow; in addition, the electronic device may further acquire a real flow from 12 o 'clock to 12 o' clock 30 minutes in 6 o 'clock in 2 th, a real flow from 12 o' clock to 12 o 'clock 30 minutes in 7 o' clock in 2 th, a real flow from 12 o 'clock to 12 o' clock 30 minutes in 8 o 'clock in 2 th, and a real flow from 12 o' clock to 12 o 'clock 30 minutes in 9 o' clock in 2 th as the longitudinal history real flow.
If the month is taken as a prediction period and the time to be predicted is 5 months and 10 days, the electronic equipment can obtain the real flow of 5 months and 6 days, the real flow of 5 months and 7 days, the real flow of 5 months and 8 days and the real flow of 5 months and 9 days as the transverse historical real flow; in addition, the electronic device may further acquire the real traffic of 1 month and 10 days, the real traffic of 2 months and 10 days, the real traffic of 3 months and 10 days, and the real traffic of 4 months and 10 days as the longitudinal history real traffic.
S102: and inputting the real flow of the time before the time to be predicted into a first pre-trained prediction network model in the prediction period to which the time to be predicted belongs to obtain the first predicted flow.
The first prediction network model may be configured to predict, according to real traffic at multiple times in the same prediction period, traffic at other times after the multiple times in the prediction period. The first predictive network model may be an ARIMA (differential Integrated Moving Average Autoregressive) model, or other network model used for predictive analysis.
In application and implementation, the electronic device may input the horizontal historical real traffic at the time to be predicted to the pre-trained first prediction network model to obtain an output result (i.e., the first predicted traffic) of the first prediction network model.
It can be understood that, when the first prediction network model is trained, both the input data and the output data are real traffic of historical time, and the relationship between the time corresponding to the output data and the time corresponding to the input data is consistent with the relationship between the time to be predicted and the transverse historical time.
In one implementation, if the day is taken as the prediction period and the time to be predicted is 2 months, 10 days, 12 points and 30 points, the lateral history time may include 2 months, 10 days, 10 points, 30 points and 11 points, 30 points and 12 points. The electronic device may obtain the flow data for a plurality of days before 10 days 2 months, and the flow data for each day may include a real flow from 10 to 10 and 30 minutes, a real flow from 10 to 30 to 11, a real flow from 11 to 30 to 11, a real flow from 30 to 12 and a real flow from 12 to 30 minutes.
The electronic device may use, as input data of the first prediction network model, the real traffic from 10 point to 10 point 30 point, the real traffic from 10 point to 11 point 30 point, the real traffic from 11 point to 11 point 30 point and the real traffic from 11 point 30 point to 12 point every day before 10 days of 2 months, and use the real traffic from 12 point to 12 point 30 point as corresponding output data to perform model training until the first prediction network model reaches a preset convergence condition, so as to obtain a trained first prediction network model.
Correspondingly, the electronic device may input the real traffic of 10 th 30 th 10 th 11 th 10 th 30 th 10 th 30 th 11 th 12 th 30 th 11 th 12 th 10 th 12 th 30 th 12 th 10 th 12 th of the first network prediction model according to the horizontal historical real traffic.
S103: and inputting the real flow of the time corresponding to the time to be predicted into a second pre-trained prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain second prediction flow.
The second prediction network model can be used for predicting the flow at the same time in other prediction periods after the multiple prediction periods according to the real flow at the same time in the multiple prediction periods. The second predictive network model may be an ARIMA model, or other network model used for predictive analysis.
In application and implementation, the electronic device may input the longitudinal historical actual traffic at the time to be predicted to the second prediction network model trained in advance, so as to obtain an output result (i.e., the second predicted traffic) of the second prediction network model.
It can be understood that, when the second prediction network model is trained, both the input data and the output data are true flows at historical times, and the relationship between the time corresponding to the output data and the time corresponding to the input data is consistent with the relationship between the time to be predicted and the longitudinal historical time.
In one implementation, if the day is used as the prediction period, the time to be predicted is 2 months, 10 days, 12 points and 30 points, and the longitudinal history time may include 2 months, 6 days, 12 points and 30 points, 2 months, 7 days, 12 points and 30 points, 2 months, 8 days, 12 points and 30 points, and 2 months, 9 days, 12 points and 30 points. The electronic device can acquire the real traffic of the same time period within a plurality of days before 2 months and 10 days, wherein the time period can be 12 o 'clock to 12 o' clock and 30 minutes, and can also be other time periods. For example, the electronic device may acquire the real flow rates of 2 month, 6 days, 10 points 30 to 11 points, and 11 points to 11 points 30, the real flow rates of 2 month, 7 days, 10 points 30 to 11 points, and 11 points to 11 points 30, the real flow rates of 2 month, 8 days, 10 points 30 to 11 points, and 11 points to 11 points 30, the real flow rates of 2 month, 9 days, 10 points 30 to 11 points, and 11 points to 11 points 30, and the real flow rates of 2 month, 10 days, 10 points 30 to 11 points, and 11 points to 11 points 30.
The electronic equipment can take the actual flow of 10 points 30 to 11 points in 6 days 2 and 6 months, the actual flow of 10 points 30 to 11 points in 7 days 2 and 7 months, the actual flow of 10 points 30 to 11 points in 8 days 2 and 8 days, and the actual flow of 10 points 30 to 11 points in 9 days 2 and 9 days as the input data of the second prediction network model, and take the actual flow of 10 points 30 to 11 points in 10 days 2 and 11 days as the corresponding output data; and (3) taking the actual flow from 11 points on 6 days of 2 months to 11 points 30 minutes, the actual flow from 11 points on 7 days of 2 months to 11 points 30 minutes, the actual flow from 11 points on 8 days of 2 months to 11 points 30 minutes and the actual flow from 11 points on 9 days of 2 months to 11 points 30 minutes as input data of the second prediction network model, and taking the actual flow from 11 points on 10 days of 2 months to 11 points 30 minutes as corresponding output data to carry out model training until the second prediction network model reaches a preset convergence condition, thereby obtaining the trained second prediction network model.
Correspondingly, the electronic device may input the real traffic from 12 o 'clock to 12 o' clock 30 minutes in 6 o 'clock in 2 o' clock, 12 o 'clock to 12 o' clock 30 minutes in 7 o 'clock in 2 o' clock, 12 o 'clock to 12 o' clock 30 minutes in 8 o 'clock in 2 o' clock, and 12 o 'clock to 12 o' clock 30 minutes in 9 o 'clock in 2 o' clock to the trained second network prediction model, and at this time, the output data (i.e., the second predicted traffic) of the second network prediction model is the traffic from 12 o 'clock to 12 o' clock 30 minutes in 10 o 'clock in 2 o' clock, which is predicted according to the longitudinal historical real traffic.
In addition, the execution order of S102 and S103 is not limited in the embodiment of the present application.
S104: and determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
In the application implementation, after the electronic device obtains the first predicted flow and the second predicted flow, the electronic device may determine the predicted flow at the time to be predicted based on the first predicted flow and the second predicted flow.
The electronic device may determine the predicted flow at the time to be predicted based on the first predicted flow and the second predicted flow in various ways, for example, the electronic device may calculate a weighted sum of the first predicted flow and the second predicted flow according to respective weights of the first predicted flow and the second predicted flow, and use the weighted sum as the predicted flow at the time to be predicted, where the weight of the first predicted flow may represent a degree of influence of the horizontal historical real flow on the flow at the time to be predicted, and the weight of the second predicted flow may represent a degree of influence of the vertical historical real flow on the flow at the time to be predicted.
Therefore, based on the flow prediction method of the embodiment of the application, the predicted flow at the time to be predicted can be determined based on the first predicted flow and the second predicted flow, that is, the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, and the accuracy of the flow prediction result can be improved.
Optionally, in order to further improve the accuracy of the flow prediction result, the electronic device may process the first predicted flow and the second predicted flow according to the regression model to obtain the predicted flow at the time to be predicted, and S104 may include the following steps:
inputting the first predicted flow and the second predicted flow into a pre-trained regression model; and taking the output data of the regression model as the predicted flow at the moment to be predicted.
The regression model (which may be referred to as a first regression model) may be a regression tree model, and the first regression model is obtained by training data using predicted traffic at a plurality of historical times output by the first prediction network model and predicted traffic at a plurality of historical times output by the second prediction network model as input data and using real traffic at a plurality of historical times as output data.
The trained first regression model can determine the respective weights of the predicted flow output by the first prediction network model and the predicted flow output by the second prediction network model.
In application implementation, the electronic device may input the first predicted flow and the second predicted flow to a first regression model trained in advance, and use output data of the first regression model as a predicted flow (which may be referred to as a target predicted flow) at a time to be predicted.
In addition, before this step, the electronic device may acquire predicted traffic of a plurality of historical times output by the first prediction network model, predicted traffic of the plurality of historical times output by the second prediction network model, and real traffic of the plurality of historical times. When the first regression model is trained, the electronic device may use the predicted traffic at each historical time output by the first prediction network model and the predicted traffic at the historical time output by the second prediction network model as input data of the first regression model, use the actual traffic at the historical time as corresponding output data, and perform model training until the first regression model reaches a preset convergence condition, so as to obtain a trained first regression model.
Correspondingly, the electronic device may input the first predicted flow and the second predicted flow to the trained first regression model, at this time, output data of the first regression model (i.e., the target predicted flow) may be a weighted sum of the first predicted flow and the second predicted flow, and at this time, respective weights of the first predicted flow and the second predicted flow are learned by the first regression model according to real flows at a plurality of historical times, so that accuracy of the obtained target predicted flow is high.
Optionally, in order to further improve the accuracy of the flow prediction result, the electronic device may further determine the predicted flow at the time to be predicted by combining the time information of the time to be predicted, and S104 may include the following steps:
and inputting the first predicted flow, the second predicted flow and the time information of the time to be predicted into a pre-trained regression model, and taking the output data of the regression model as the predicted flow of the time to be predicted.
The time information of the time to be predicted may include an identifier capable of indicating a time corresponding to the time to be predicted, for example, taking a day as a prediction cycle, and dividing 24 hours in a day into 48 time periods, where the time periods are: 0 point to 0 point 30 minutes, 0 point 30 minutes to 1 point, 1 point to 1 point 30 minutes, …, 22 point 30 minutes to 23 points, 23 points to 23 points 30 minutes and 23 points 30 minutes to 24 points. The time information of the time periods may be represented by 1, 2, 3, …, 46, 47, and 48, respectively. The time information may also include the date of the day of the time to be predicted, for example, the day of the time to be predicted is 10 of a month, and the time information of the time to be predicted may include 10, which is used to indicate that the day of the time to be predicted is 10; if the day of the time to be predicted is friday, the time information of the time to be predicted can also comprise 5 which is used for indicating that the day of the time to be predicted is friday, and in the actual operation process, technicians can set the identifier contained in the time information according to business requirements.
The regression model (may be referred to as a second regression model) may be a regression tree model, and the second regression model may be obtained by training using the predicted traffic at the plurality of historical times output by the first prediction network model, the predicted traffic at the plurality of historical times output by the second prediction network model, and the time information at the plurality of historical times as input data and using the actual traffic at the plurality of historical times as output data.
In application implementation, the electronic device may input the first predicted flow, the second predicted flow, and the time information at the time to be predicted to a second regression model trained in advance, and use output data of the second regression model as the predicted flow (i.e., the target predicted flow) at the time to be predicted.
In addition, before this step, the electronic device may acquire predicted traffic volumes of a plurality of historical times output by the first prediction network model, predicted traffic volumes of the plurality of historical times output by the second prediction network model, actual traffic volumes of the plurality of historical times, and time information of the plurality of historical times. When the second regression model is trained, the electronic device may use the predicted flow at each historical time output by the first prediction network model, the predicted flow at the historical time output by the second prediction network model, and the time information at the historical time as input data of the second regression model, use the actual flow at the historical time as corresponding output data, and perform model training until the second regression model reaches a preset convergence condition, so as to obtain a trained second regression model.
Correspondingly, the electronic device may input the first predicted flow, the second predicted flow and the time information of the time to be predicted to the trained second regression model, and at this time, the output data (i.e., the target predicted flow) of the second regression model is the flow determined by combining the first predicted flow, the second predicted flow and the time information of the time to be predicted, so that the accuracy of the obtained target predicted flow is high.
In addition, in order to improve the accuracy of the flow prediction result, the electronic device may further determine whether the flow at the time to be predicted needs to be calculated according to the horizontal history actual flow and the vertical history actual flow according to the time lengths of the time to be predicted and the current time, and before S101, the method may further include the following steps:
judging whether the distance between the moment to be predicted and the current moment is greater than a preset duration or not; and if the distance between the moment to be predicted and the current moment is not more than the preset time length, executing S101.
In the application embodiment, the electronic device may determine whether the distance between the time to be predicted and the current time is greater than a preset duration, and when the electronic device determines that the distance between the time to be predicted and the current time is not greater than the preset duration, the electronic device may obtain horizontal historical real traffic and vertical historical real traffic of the time to be predicted, and predict the traffic of the time to be predicted according to the first prediction network model and the second prediction network model.
The preset time duration may be set by a technician according to experience, for example, the preset time duration is 3 hours, a day is used as a prediction period, if the current time is 9 points at 10 months and 10 days, the time to be predicted is 10 points at 10 days and 30 minutes, and the time to be predicted is 1 hour away from the current time and is less than the preset time duration, the electronic device may obtain the horizontal historical real flow and the vertical historical real flow at the time to be predicted, so as to predict the flow at the time to be predicted.
Optionally, the method may further include the steps of:
step one, if the distance between the time to be predicted and the current time is longer than the preset time length, acquiring the real flow of the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs.
In the application implementation, when the electronic device determines that the time to be predicted is greater than the preset duration from the current time, at this time, the real traffic does not exist at the time closer to the time to be predicted, that is, the electronic device cannot acquire the horizontal historical real traffic of the time to be predicted, and the electronic device can acquire the longitudinal historical real traffic of the time to be predicted so as to predict the traffic of the time to be predicted according to the longitudinal historical real traffic.
And step two, inputting the real flow of the time corresponding to the time to be predicted into a second pre-trained prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
In the application implementation, after the electronic device obtains the longitudinal historical actual traffic at the time to be predicted, the electronic device may input the longitudinal historical actual traffic at the time to be predicted to the trained second prediction network model, and use output data (i.e., the second predicted traffic) of the second prediction network model as the predicted traffic at the time to be predicted.
Based on the above processing, if the distance between the time to be predicted and the current time is longer than the preset time, it is indicated that the electronic device cannot acquire the transverse historical real flow of the time to be predicted, at this time, the electronic device can predict the flow of the time to be predicted according to the longitudinal historical real flow of the time to be predicted, and the situation that the flow cannot be predicted is avoided to a certain extent.
Referring to fig. 2, fig. 2 is a flowchart of an example of a flow prediction method provided in an embodiment of the present application, where the method may include the following steps:
s201: and judging whether the distance between the moment to be predicted and the current moment is greater than a preset time length or not, if the distance between the moment to be predicted and the current moment is not greater than the preset time length, executing S202-S206, and if the distance between the moment to be predicted and the current moment is greater than the preset time length, executing S207-S208.
S202: and acquiring the real flow of the time before the time to be predicted in the prediction period to which the time to be predicted belongs, and the real flow of the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs.
S203: and inputting the real flow of the time before the time to be predicted into a first pre-trained prediction network model in the prediction period to which the time to be predicted belongs to obtain the first predicted flow.
S204: and inputting the real flow of the time corresponding to the time to be predicted into a second pre-trained prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain second prediction flow.
S205: and inputting the first predicted flow, the second predicted flow and the time information of the moment to be predicted into a pre-trained regression model.
The regression model is obtained by training by taking the predicted flow at a plurality of historical moments output by the first prediction network model, the predicted flow at the plurality of historical moments output by the second prediction network model and the time information at the plurality of historical moments as input data and taking the real flow at the plurality of historical moments as output data.
S206: and taking the output data of the regression model as the predicted flow at the moment to be predicted.
S207: and acquiring the real flow of the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs.
S208: and inputting the real flow of the time corresponding to the time to be predicted into a second pre-trained prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
Therefore, based on the flow prediction method provided by the embodiment of the application, the predicted flow at the time to be predicted can be determined based on the first predicted flow and the second predicted flow, that is, the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, and the accuracy of the flow prediction result can be improved.
Corresponding to the embodiment of the method in fig. 1, referring to fig. 3, fig. 3 is a block diagram of a flow prediction apparatus provided in an embodiment of the present application, where the apparatus may include:
an obtaining module 301, configured to obtain a real flow at a time before a time to be predicted in a prediction period to which the time to be predicted belongs, and a real flow at a time corresponding to the time to be predicted in a prediction period before the prediction period to which the time to be predicted belongs;
a first prediction module 302, configured to input, in a prediction period to which the time to be predicted belongs, a real traffic at a time before the time to be predicted to a first pre-trained prediction network model to obtain a first predicted traffic, where the first prediction network model is configured to predict, according to real traffic at multiple times in the same prediction period, traffic at other times after the multiple times in the prediction period;
a second prediction module 303, configured to input, to a second pre-trained prediction network model, the actual traffic at the time corresponding to the time to be predicted in a prediction period before a prediction period to which the time to be predicted belongs, so as to obtain a second predicted traffic, where the second prediction network model is configured to predict, according to the actual traffic at the same time in multiple prediction periods, the traffic at the same time in other prediction periods after the multiple prediction periods;
a determining module 304, configured to determine the predicted flow at the time to be predicted based on the first predicted flow and the second predicted flow.
Optionally, the determining module 304 is specifically configured to input the first predicted traffic and the second predicted traffic to a regression model trained in advance, where the regression model is obtained by training using predicted traffic at a plurality of historical times output by the first predicted network model and predicted traffic at a plurality of historical times output by the second predicted network model as input data, and using real traffic at a plurality of historical times as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
Optionally, the determining module 304 is specifically configured to input the first predicted traffic, the second predicted traffic, and the time information of the time to be predicted into a regression model trained in advance, where the regression model is obtained by training using predicted traffic of a plurality of historical times output by the first predicted network model, predicted traffic of a plurality of historical times output by the second predicted network model, and the time information of the plurality of historical times as input data, and using real traffic of the plurality of historical times as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
Optionally, the apparatus further comprises:
the judging module is used for judging whether the distance between the moment to be predicted and the current moment is greater than a preset duration;
if the distance between the time to be predicted and the current time is not more than the preset time, the obtaining module 301 is triggered.
Optionally, the apparatus further comprises:
the processing module is used for acquiring the real flow of the time corresponding to the time to be predicted in a prediction period before the prediction period to which the time to be predicted belongs if the distance between the time to be predicted and the current time is greater than the preset time length;
and inputting the real flow of the time corresponding to the time to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
As can be seen from the above, based on the traffic prediction apparatus provided in the embodiment of the present application, the real traffic at the time before the time to be predicted in the prediction period to which the time to be predicted belongs may be input to the first pre-trained prediction network model to obtain the first predicted traffic, the real traffic at the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs may be input to the second pre-trained prediction network model to obtain the second predicted traffic, and the predicted traffic at the time to be predicted may be determined based on the first predicted traffic and the second predicted traffic. In the prior art, the first predicted flow is used as the predicted flow at the time to be predicted, and the predicted flow at the time to be predicted is determined based on the first predicted flow and the second predicted flow, that is, the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, so that the accuracy of the flow prediction result can be improved.
The embodiment of the present application further provides an electronic device, as shown in fig. 4, which includes a processor 401, a communication interface 402, a memory 403, and a communication bus 404, where the processor 401, the communication interface 402, and the memory 403 complete mutual communication through the communication bus 404,
a memory 403 for storing a computer program;
the processor 401, when executing the program stored in the memory 403, implements the following steps:
acquiring real flow of a moment before the moment to be predicted in a prediction period to which the moment to be predicted belongs, and real flow of a moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
inputting the real traffic of the time before the time to be predicted into a pre-trained first prediction network model in a prediction cycle to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction cycle according to the real traffic of the times in the same prediction cycle;
inputting the real traffic of the moment corresponding to the moment to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to obtain second prediction traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
The communication bus mentioned in the electronic device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface is used for communication between the electronic equipment and other equipment.
The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
When the electronic device provided by the embodiment of the application performs flow prediction, the predicted flow at the time to be predicted can be determined based on the first predicted flow and the second predicted flow, that is, the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, so that the accuracy of a flow prediction result can be improved.
The embodiment of the present application further provides a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer is enabled to execute the flow prediction method provided by the embodiment of the present application.
Specifically, the traffic prediction method includes:
acquiring real flow of a moment before the moment to be predicted in a prediction period to which the moment to be predicted belongs, and real flow of a moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
inputting the real traffic of the time before the time to be predicted into a pre-trained first prediction network model in a prediction cycle to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction cycle according to the real traffic of the times in the same prediction cycle;
inputting the real traffic of the moment corresponding to the moment to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to obtain second prediction traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
It should be noted that other implementation manners of the traffic prediction method are the same as those of the foregoing method embodiment, and are not described herein again.
By operating the instructions stored in the computer-readable storage medium provided by the embodiment of the application, when flow prediction is performed, the predicted flow at the time to be predicted can be determined based on the first predicted flow and the second predicted flow, that is, the influence of real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, and the accuracy of a flow prediction result can be improved.
Embodiments of the present application further provide another computer program product containing instructions, which when executed on a computer, cause the computer to execute the flow prediction method provided by embodiments of the present application.
Specifically, the traffic prediction method includes:
acquiring real flow of a moment before the moment to be predicted in a prediction period to which the moment to be predicted belongs, and real flow of a moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
inputting the real traffic of the time before the time to be predicted into a pre-trained first prediction network model in a prediction cycle to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction cycle according to the real traffic of the times in the same prediction cycle;
inputting the real traffic of the moment corresponding to the moment to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to obtain second prediction traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
It should be noted that other implementation manners of the traffic prediction method are the same as those of the foregoing method embodiment, and are not described herein again.
By operating the computer program product provided by the embodiment of the application, when the flow is predicted, the predicted flow at the time to be predicted can be determined based on the first predicted flow and the second predicted flow, that is, the influence of the real flow at the time corresponding to the time to be predicted in different prediction periods on the flow at the time to be predicted is comprehensively considered, so that the accuracy of the flow prediction result can be improved.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. 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, cause the processes or functions described in accordance with the embodiments of the application to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus, the electronic device, the computer-readable storage medium, and the computer program product embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiments.
The above description is only for the preferred embodiment of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application are included in the protection scope of the present application.
Claims (12)
1. A method of traffic prediction, the method comprising:
acquiring real flow of a moment before the moment to be predicted in a prediction period to which the moment to be predicted belongs, and real flow of a moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
inputting the real traffic of the time before the time to be predicted into a pre-trained first prediction network model in a prediction cycle to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction cycle according to the real traffic of the times in the same prediction cycle;
inputting the real traffic of the moment corresponding to the moment to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to obtain second prediction traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
2. The method of claim 1, wherein the determining the predicted flow rate at the time to be predicted based on the first predicted flow rate and the second predicted flow rate comprises:
inputting the first predicted flow and the second predicted flow to a pre-trained regression model, wherein the regression model is obtained by training by taking the predicted flows of a plurality of historical moments output by the first predicted network model and the predicted flows of the plurality of historical moments output by the second predicted network model as input data and taking the real flows of the plurality of historical moments as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
3. The method of claim 1, wherein the determining the predicted flow rate at the time to be predicted based on the first predicted flow rate and the second predicted flow rate comprises:
inputting the first predicted flow, the second predicted flow and the time information of the time to be predicted into a regression model trained in advance, wherein the regression model is obtained by training by taking the predicted flows of a plurality of historical moments output by the first predicted network model, the predicted flows of the plurality of historical moments output by the second predicted network model and the time information of the plurality of historical moments as input data and the real flows of the plurality of historical moments as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
4. The method according to claim 1, wherein before the obtaining of the real flow at a time before the time to be predicted in the prediction period to which the time to be predicted belongs and the real flow at a time corresponding to the time to be predicted in the prediction period to which the time to be predicted belongs, the method further comprises:
judging whether the distance between the moment to be predicted and the current moment is greater than a preset duration or not;
and if the distance between the moment to be predicted and the current moment is not more than the preset time length, executing the step of acquiring the real flow of the moment to be predicted in the prediction period before the moment to be predicted and the real flow of the moment corresponding to the moment to be predicted in the prediction period before the moment to be predicted.
5. The method of claim 4, further comprising:
if the distance between the moment to be predicted and the current moment is greater than the preset duration, acquiring the real flow of the moment corresponding to the moment to be predicted in a prediction period before the prediction period to which the moment to be predicted belongs;
and inputting the real flow of the time corresponding to the time to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
6. A flow prediction apparatus, characterized in that the apparatus comprises:
the acquisition module is used for acquiring the real flow of the time before the time to be predicted in the prediction period to which the time to be predicted belongs and the real flow of the time corresponding to the time to be predicted in the prediction period before the prediction period to which the time to be predicted belongs;
the first prediction module is used for inputting the real traffic of the time before the time to be predicted into a first pre-trained prediction network model in a prediction period to which the time to be predicted belongs to obtain first predicted traffic, wherein the first prediction network model is used for predicting the traffic of other times after the times in the prediction period according to the real traffic of the times in the same prediction period;
the second prediction module is used for inputting the real traffic of the moment corresponding to the moment to be predicted into a second pre-trained prediction network model in a prediction period before the prediction period to which the moment to be predicted belongs to so as to obtain second predicted traffic, wherein the second prediction network model is used for predicting the traffic of the same moment in other prediction periods after the multiple prediction periods according to the real traffic of the same moment in the multiple prediction periods;
and the determining module is used for determining the predicted flow at the moment to be predicted based on the first predicted flow and the second predicted flow.
7. The apparatus according to claim 6, wherein the determining module is specifically configured to input the first predicted flow and the second predicted flow into a regression model trained in advance, where the regression model is obtained by training using predicted flows at a plurality of historical times output by the first prediction network model and predicted flows at a plurality of historical times output by the second prediction network model as input data and using real flows at a plurality of historical times as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
8. The apparatus according to claim 6, wherein the determining module is specifically configured to input the first predicted traffic, the second predicted traffic, and the time information of the time to be predicted into a regression model trained in advance, where the regression model is obtained by using predicted traffic of a plurality of historical times output by the first predicted network model, predicted traffic of the plurality of historical times output by the second predicted network model, and the time information of the plurality of historical times as input data and training using real traffic of the plurality of historical times as output data;
and taking the output data of the regression model as the predicted flow at the moment to be predicted.
9. The apparatus of claim 6, further comprising:
the judging module is used for judging whether the distance between the moment to be predicted and the current moment is greater than a preset duration;
and if the distance between the moment to be predicted and the current moment is not more than the preset time length, triggering the acquisition module.
10. The apparatus of claim 9, further comprising:
the processing module is used for acquiring the real flow of the time corresponding to the time to be predicted in a prediction period before the prediction period to which the time to be predicted belongs if the distance between the time to be predicted and the current time is greater than the preset time length;
and inputting the real flow of the time corresponding to the time to be predicted into a pre-trained second prediction network model in a prediction period before the prediction period to which the time to be predicted belongs to, so as to obtain the predicted flow of the time to be predicted.
11. An electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;
the memory is used for storing a computer program;
the processor, when executing the program stored in the memory, implementing the method steps of any of claims 1-5.
12. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of the claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910305627.0A CN111835536B (en) | 2019-04-16 | 2019-04-16 | Flow prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910305627.0A CN111835536B (en) | 2019-04-16 | 2019-04-16 | Flow prediction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111835536A true CN111835536A (en) | 2020-10-27 |
| CN111835536B CN111835536B (en) | 2022-04-05 |
Family
ID=72915226
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910305627.0A Active CN111835536B (en) | 2019-04-16 | 2019-04-16 | Flow prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111835536B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343956A (en) * | 2021-08-06 | 2021-09-03 | 腾讯科技（深圳）有限公司 | Road condition information prediction method and device, storage medium and electronic equipment |
| CN113556253A (en) * | 2021-07-30 | 2021-10-26 | 济南浪潮数据技术有限公司 | Method, system, device and storage medium for predicting real-time flow of switch port |
| CN113888060A (en) * | 2021-11-24 | 2022-01-04 | 中国工商银行股份有限公司 | Method, apparatus, electronic device and medium for determining website operation strategy |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102360388A (en) * | 2011-10-20 | 2012-02-22 | 苏州大学 | Time series forecasting method and system based on SVR (Support Vector Regression) |
| CN104993966A (en) * | 2015-07-15 | 2015-10-21 | 国家电网公司 | Power integrated service network flow prediction method |
| CN105355038A (en) * | 2015-10-14 | 2016-02-24 | 青岛观澜数据技术有限公司 | Method for predicting short-term traffic flow through employing PMA modeling |
| CN108345857A (en) * | 2018-02-09 | 2018-07-31 | 北京天元创新科技有限公司 | A kind of region crowd density prediction technique and device based on deep learning |
| US20180302296A1 (en) * | 2016-10-14 | 2018-10-18 | Tencent Technology (Shenzhen) Company Limited | Network service scheduling method and apparatus, storage medium, and program product |
-
2019
- 2019-04-16 CN CN201910305627.0A patent/CN111835536B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102360388A (en) * | 2011-10-20 | 2012-02-22 | 苏州大学 | Time series forecasting method and system based on SVR (Support Vector Regression) |
| CN104993966A (en) * | 2015-07-15 | 2015-10-21 | 国家电网公司 | Power integrated service network flow prediction method |
| CN105355038A (en) * | 2015-10-14 | 2016-02-24 | 青岛观澜数据技术有限公司 | Method for predicting short-term traffic flow through employing PMA modeling |
| US20180302296A1 (en) * | 2016-10-14 | 2018-10-18 | Tencent Technology (Shenzhen) Company Limited | Network service scheduling method and apparatus, storage medium, and program product |
| CN108345857A (en) * | 2018-02-09 | 2018-07-31 | 北京天元创新科技有限公司 | A kind of region crowd density prediction technique and device based on deep learning |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113556253A (en) * | 2021-07-30 | 2021-10-26 | 济南浪潮数据技术有限公司 | Method, system, device and storage medium for predicting real-time flow of switch port |
| CN113556253B (en) * | 2021-07-30 | 2023-05-26 | 济南浪潮数据技术有限公司 | Method, system, equipment and storage medium for predicting real-time traffic of switch port |
| CN113343956A (en) * | 2021-08-06 | 2021-09-03 | 腾讯科技（深圳）有限公司 | Road condition information prediction method and device, storage medium and electronic equipment |
| CN113888060A (en) * | 2021-11-24 | 2022-01-04 | 中国工商银行股份有限公司 | Method, apparatus, electronic device and medium for determining website operation strategy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111835536B (en) | 2022-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107480028B (en) | Method and device for acquiring usable residual time of disk | |
| CN107040397B (en) | Service parameter acquisition method and device | |
| CN109190808B (en) | User behavior prediction method, device, equipment and medium | |
| CN111835536B (en) | Flow prediction method and device | |
| CN108985489B (en) | Risk prediction method, risk prediction device and terminal equipment | |
| CN112882889A (en) | Abnormality monitoring method, abnormality monitoring system, electronic device, and storage medium | |
| CN108243032B (en) | Method, device and equipment for acquiring service level information | |
| WO2020253038A1 (en) | Model construction method and apparatus | |
| CN109003091A (en) | A kind of risk prevention system processing method, device and equipment | |
| CN111047429A (en) | Probability prediction method and device | |
| CN112669078A (en) | Behavior prediction model training method, device, equipment and storage medium | |
| CN109685255A (en) | A kind of method and apparatus for predicting customer churn | |
| CN111784160A (en) | River hydrological situation change evaluation method and system | |
| CN108805332B (en) | Feature evaluation method and device | |
| CN111740865B (en) | Flow fluctuation trend prediction method and device and electronic equipment | |
| CN113378498B (en) | Task allocation method and device | |
| CN111800807A (en) | Method and device for alarming number of base station users | |
| CN113123955B (en) | Plunger pump abnormity detection method and device, storage medium and electronic equipment | |
| CN110505635B (en) | Terminal permeability prediction method and device | |
| JP2015184818A (en) | Server, model application propriety determination method and computer program | |
| CN116011677A (en) | Time sequence data prediction method and device, electronic equipment and storage medium | |
| CN113537631B (en) | Medicine demand prediction method, device, electronic equipment and storage medium | |
| CN113657635B (en) | Method for predicting loss of communication user and electronic equipment | |
| CN102916851A (en) | Network traffic prediction method and network traffic prediction device | |
| CN113129127A (en) | Early warning method and device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |