Traffic prediction method, device, equipment, storage medium and program product
By combining the Prophet model and the N-BEATS model, the network traffic data is decomposed into trend and random terms and the residuals are corrected, which solves the problem of low accuracy in network traffic prediction, achieves higher accuracy in traffic prediction, and supports network planning optimization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA TELECOM CORP LTD TECHNOLOGY INNOVATION CENTER
- Filing Date
- 2023-07-07
- Publication Date
- 2026-07-31
AI Technical Summary
Existing network traffic prediction methods are not accurate enough to meet the needs of network planning.
By acquiring network traffic data from the first region and the second region that influences it, the Prophet model is used to decompose and predict the trend and random terms, and the N-BEATS model is combined to correct the residuals. By comprehensively considering the mutual influence between regions, the prediction accuracy is improved.
It improves the accuracy of network traffic prediction, enabling more accurate prediction of network traffic trends and scale, and supports network structure optimization and capacity expansion.
Smart Images
Figure CN116781536B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of network communication technology, and in particular to a traffic prediction method, apparatus, device, storage medium, and program product. Background Technology
[0002] To adapt to the rapid growth of network traffic and meet the needs of upper-layer services and applications, operators need to closely monitor the current state of network traffic, promptly assess and predict the trend and scale of network traffic during the planning period, and use this as input for subsequent network planning, ultimately achieving optimization and transformation of network structure and bandwidth distribution or capacity expansion. Therefore, accurate prediction of network traffic is of great significance to network planning.
[0003] However, current traditional network traffic prediction methods do not have high accuracy in predicting network traffic. Summary of the Invention
[0004] Therefore, it is necessary to provide a flow prediction method, apparatus, device, storage medium, and program product that can improve prediction accuracy in response to the above-mentioned technical problems.
[0005] In a first aspect, this application provides a traffic prediction method. The method includes: acquiring first network traffic data within a first historical time period, the first historical time period preceding a target time period to be predicted; the first network traffic data including network traffic data of a first region and network traffic data of a second region, wherein the second region is the region where network traffic affects network traffic in the first region; inputting the first network traffic data into a first prediction model to obtain first predicted traffic data for the first region within the target time period; inputting the historical residual corresponding to the first historical time period into a second prediction model to obtain prediction residual corresponding to the target time period, the historical residual being used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical time period and the actual network traffic data of the first region in the first historical time period; and determining the predicted traffic data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0006] In one embodiment, before acquiring the first network traffic data within a first historical time period, the method further includes: calculating the correlation between the network traffic of the first region and the network traffic of other multiple candidate regions; and determining a second region from the multiple candidate regions based on the calculated correlation.
[0007] In one embodiment, calculating the correlation between network traffic in the first region and network traffic in other multiple candidate regions includes: for each candidate region, calculating the covariance between network traffic data in the candidate region and network traffic data in the first region, calculating the first variance of network traffic data in the candidate region, and calculating the second variance of network traffic data in the first region, and calculating the correlation between network traffic in the first region and network traffic in the candidate regions based on the covariance, the first variance, and the second variance.
[0008] In one embodiment, inputting first network traffic data into a first prediction model to obtain first predicted traffic data for a first region within a target time period includes: inputting the first network traffic data into the first prediction model, decomposing the first network traffic data into trend items and random items through the first prediction model, performing traffic prediction processing on the trend items and random items respectively, and fusing the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0009] In one embodiment, before inputting the historical residual corresponding to the first historical time period into the second prediction model, the method further includes: acquiring second network traffic data within the second historical time period, wherein the second historical time period is prior to the first historical time period, and the second network traffic data includes network traffic data of the first region and network traffic data of the second region; inputting the second network traffic data into the first prediction model to obtain second predicted traffic data of the first region within the first historical time period; and determining the historical residual based on the difference between the second predicted traffic data and the actual network traffic data of the first region in the first historical time period.
[0010] In one embodiment, determining the traffic prediction data for a first region within a target time period based on the first predicted traffic data and the prediction residual includes: correcting the first predicted traffic data according to the prediction residual to obtain the traffic prediction data.
[0011] In one embodiment, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
[0012] Secondly, this application also provides a flow prediction device. The device includes:
[0013] The acquisition module is used to acquire the first network traffic data within the first historical time period. The first historical time period is located before the target time period to be predicted. The first network traffic data includes network traffic data of the first region and network traffic data of the second region. The second region is the region where network traffic affects network traffic in the first region.
[0014] The first prediction module is used to input the first network traffic data into the first prediction model to obtain the first predicted traffic data of the first region within the target time period;
[0015] The second prediction module is used to input the historical residual corresponding to the first historical period into the second prediction model to obtain the prediction residual corresponding to the target period. The historical residual is used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical period and the actual network traffic data of the first region in the first historical period.
[0016] The determination module is used to determine the traffic forecast data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0017] In one embodiment, the apparatus further includes a computing module for calculating the correlation between network traffic in the first region and network traffic in other multiple candidate regions; and determining a second region from the multiple candidate regions based on the calculated correlation.
[0018] In one embodiment, the calculation module is specifically used to calculate, for each candidate region, the covariance between the network traffic data of the candidate region and the network traffic data of the first region, calculate the first variance of the network traffic data of the candidate region, and calculate the second variance of the network traffic data of the first region, and calculate the correlation between the network traffic of the first region and the network traffic of the candidate region based on the covariance, the first variance and the second variance.
[0019] In one embodiment, the first prediction module is specifically used to input the first network traffic data into the first prediction model, decompose the first network traffic data into trend items and random items through the first prediction model, perform traffic prediction processing on the trend items and random items respectively, and fuse the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0020] In one embodiment, the device further includes a residual determination module for acquiring second network traffic data within a second historical time period, the second historical time period being prior to the first historical time period, the second network traffic data including network traffic data of a first region and network traffic data of a second region; inputting the second network traffic data into a first prediction model to obtain second predicted traffic data of the first region within the first historical time period; and determining historical residuals based on the difference between the second predicted traffic data and the actual network traffic data of the first region within the first historical time period.
[0021] In one embodiment, the determining module is specifically configured to correct the first predicted flow data based on the prediction residual to obtain flow prediction data.
[0022] In one embodiment, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
[0023] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the traffic prediction method described in any of the first aspects above.
[0024] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the traffic prediction method described in any of the first aspects above.
[0025] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the traffic prediction method described in any of the first aspects above.
[0026] The aforementioned traffic prediction method, apparatus, device, storage medium, and program product first acquire first network traffic data within a first historical time period preceding the target time period to be predicted. This first network traffic data includes network traffic data from a first region and network traffic data from a second region that influences the network traffic in the first region. Then, the first network traffic data is input into a first prediction model to obtain the first predicted traffic data for the first region within the target time period. The historical residuals corresponding to the first historical time period are input into a second prediction model to obtain the prediction residuals for the target time period. Finally, the predicted traffic data for the first region within the target time period is determined based on the first predicted traffic data and the prediction residuals. The historical residuals characterize the difference between the network traffic data for the first region predicted by the first prediction model in the first historical time period and the actual network traffic data for the first region in the first historical time period. In this way, the mutual influence of network traffic between the first region and the second region is comprehensively considered in the traffic prediction process. The first network traffic data, which includes network traffic data from the first region and the second region, is predicted by the first prediction model to obtain the first predicted traffic data. Then, the prediction residual of the first prediction model is obtained by the second prediction model based on the historical residual. Finally, the first predicted traffic data is corrected by the prediction residual to obtain the traffic prediction data for the target time period, so that the traffic prediction results are more accurate. Attached Figure Description
[0027] Figure 1 This is a flowchart illustrating a traffic prediction method in one embodiment;
[0028] Figure 2 This is a flowchart illustrating the traffic prediction method in another embodiment;
[0029] Figure 3 This is a flowchart illustrating the historical residual determination step in another embodiment;
[0030] Figure 4 This is a schematic diagram of the N-BEATS model in another embodiment;
[0031] Figure 5 This is a flowchart illustrating the traffic prediction method in another embodiment;
[0032] Figure 6 This is a structural block diagram of a flow prediction device in one embodiment;
[0033] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0034] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0035] In one embodiment, such as Figure 1 As shown, a traffic prediction method is provided. Taking its application to a terminal as an example, it can be understood that this method can also be applied to a server, and to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The method includes the following steps:
[0036] Step 101: Obtain the first network traffic data within the first historical time period.
[0037] The first historical time period is located before the target time period to be predicted. The first network traffic data includes network traffic data from a first region and network traffic data from a second region, where the second region is the area where network traffic affects the network traffic of the first region. Optionally, the first network traffic data is a time series; the first region can be a specific province, and the second region can be other provinces where network traffic affects the network traffic of the first region, meaning there is a certain correlation between the network traffic in the second region and the network traffic in the first region. By obtaining the first network traffic data of the first historical time period before the target time period to be predicted, it is possible to predict the network traffic for the target time period.
[0038] Step 102: Input the first network traffic data into the first prediction model to obtain the first predicted traffic data of the first region within the target time period.
[0039] Optionally, the first prediction model can be trained from the Prophet model, which is capable of time series prediction. The first prediction model can also be other machine learning algorithms capable of time series prediction, such as Recurrent Neural Networks (RNNs) or Long Short-Term Memory (LSTMs). Optionally, the initial Prophet model can be trained using historical network traffic data, and its parameters adjusted to obtain a first prediction model with optimal parameters. Inputting the first network traffic data into the trained first prediction model yields the first predicted traffic data for the first region within the target time period.
[0040] Step 103: Input the historical residuals corresponding to the first historical period into the second prediction model to obtain the prediction residuals corresponding to the target period.
[0041] The historical residual is used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical period and the actual network traffic data of the first region in the first historical period. The second prediction model can be trained based on the N-BEATS model, or it can be trained by other models capable of predicting time series data. Optionally, the initial N-BEATS model can be trained using the historical residual data of network traffic data in historical periods to obtain the second prediction model with optimal parameters.
[0042] Step 104: Determine the traffic forecast data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0043] The prediction residual can characterize the residual in the first predicted flow data obtained by the first prediction model for the target period. Therefore, the first predicted flow data can be corrected by the prediction residual. Optionally, the sum of the first predicted flow data and the prediction residual can be used as the flow prediction data of the first region in the target period.
[0044] In the above embodiments, firstly, first network traffic data within a first historical time period preceding the target time period to be predicted is acquired. This first network traffic data includes network traffic data from a first region and network traffic data from a second region that influences the network traffic in the first region. Then, the first network traffic data is input into a first prediction model to obtain the first predicted traffic data for the first region within the target time period. The historical residuals corresponding to the first historical time period are input into a second prediction model to obtain the prediction residuals for the target time period. Finally, the predicted traffic data for the first region within the target time period is determined based on the first predicted traffic data and the prediction residuals. The historical residuals characterize the difference between the network traffic data predicted by the first prediction model for the first region during the first historical time period and the actual network traffic data for the first region during the first historical time period. In this way, the mutual influence of network traffic between the first and second regions is comprehensively considered during the traffic prediction process. The first network traffic data, which includes network traffic data from both regions, is predicted using the first prediction model to obtain the first predicted traffic data. Then, the second prediction model predicts based on the historical residuals to obtain the prediction residuals. Finally, the prediction residuals are used to correct the first predicted traffic data to obtain the predicted traffic data for the target time period, resulting in higher accuracy in the predicted traffic data.
[0045] In one embodiment, before obtaining the first network traffic data within the first historical time period, it is also necessary to first determine the second region that affects the network traffic of the first region, such as... Figure 2 As shown, the steps for determining the second region include:
[0046] Step 201: Calculate the correlation between the network traffic of the first region and the network traffic of other candidate regions.
[0047] Correlation refers to the degree to which network traffic from other candidate regions affects network traffic from the first region.
[0048] Optionally, the correlation calculation process is as follows: For each candidate region, calculate the covariance between the network traffic data of the candidate region and the network traffic data of the first region; calculate the first variance of the network traffic data of the candidate region; and calculate the second variance of the network traffic data of the first region. Based on the covariance, the first variance, and the second variance, calculate the correlation between the network traffic of the first region and the network traffic of the candidate region. The calculation formula is as follows:
[0049]
[0050] Where X represents the network traffic data of the first region, Y represents the network traffic data of a candidate region, Cov|X,Y| is the covariance between the network traffic data of the candidate region and the network traffic data of the first region, Var|Y| is the first variance of the network traffic data of the candidate region, and Var|X| is the second variance of the network traffic data of the first region.
[0051] Step 202: Determine the second region from multiple candidate regions based on the calculated correlation.
[0052] The correlation between the first region and other candidate regions is calculated using the formula above. The correlation is a scaled representation of the covariance between -1 and 1, where 1 represents a strong positive correlation, 0 represents independence, and -1 represents a strong negative correlation. After calculating the correlation between the first region and multiple candidate regions, the absolute values of all correlation calculations are taken, and the candidate regions are sorted according to their absolute correlation values from largest to smallest. Optionally, a preset number of candidate regions with the highest absolute correlation values are selected as the second region. If the preset number is 3, then the top 3 candidate regions are selected as the second region.
[0053] In the above embodiments, by calculating the correlation between the first region and the candidate regions, a second region with a high correlation to the first region is determined from multiple candidate regions. This takes into account the mutual influence between different regions and can improve the accuracy of traffic prediction. Furthermore, using the network traffic data of the first region and the network traffic data of the second region as the first network traffic data enriches the amount of first network traffic data and further improves the accuracy of traffic prediction.
[0054] In the embodiments of this application, in order to improve the prediction accuracy, the step of obtaining the first predicted traffic data of the first region within the target time period according to the first prediction model includes: inputting the first network traffic data into the first prediction model, decomposing the first network traffic data into trend items and random items through the first prediction model, performing traffic prediction processing on the trend items and random items respectively, and fusing the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0055] Optionally, based on the characteristics of the first network traffic data, the first network traffic data can be decomposed into the following formula using the first prediction model:
[0056] y(t) = g(t) + r(t)
[0057] Where g(t) is the trend term and r(t) is the random term.
[0058] The first prediction model is used to perform flow prediction processing on the trend item and the random item respectively. Then, the flow prediction processing results of the trend item and the flow prediction processing results of the random item are added together to obtain the first predicted flow data.
[0059] In the above embodiments, by decomposing the first network traffic, predicting it separately, and then fusing the results, the trend items and random items with different characteristics can be predicted separately, thereby improving the accuracy of traffic prediction.
[0060] In one embodiment, the second prediction model predicts residuals based on historical residuals, such as... Figure 3 As shown, the process of determining historical residuals includes:
[0061] Step 301: Obtain the second network traffic data within the second historical time period.
[0062] The second historical period is located before the first historical period, and the second network traffic data includes network traffic data from the first region and network traffic data from the second region.
[0063] Step 302: Input the second network traffic data into the first prediction model to obtain the second predicted traffic data for the first region within the first historical period.
[0064] The first prediction model can make predictions based on the second network traffic data of the second historical period to obtain the second predicted traffic data of the first region within the first historical period.
[0065] Step 303: Determine the historical residual based on the difference between the second predicted traffic data and the actual network traffic data of the first region in the first historical period.
[0066] Optionally, the actual network traffic data for the first historical period can be obtained from historical data, and then the difference between the second predicted traffic data obtained by predicting the first historical period and the actual network traffic data for the first historical period can be used as the historical residual for the first historical period.
[0067] By inputting the historical residuals of the first historical period into the second prediction model, the prediction residuals for the target period can be obtained.
[0068] In one embodiment, determining traffic forecast data for a first region within a target time period based on first predicted traffic data and prediction residuals includes: correcting the first predicted traffic data according to the prediction residuals to obtain the traffic forecast data.
[0069] Optionally, the predicted residuals and the first predicted flow data can be summed to obtain the flow prediction data for the first region during the target period.
[0070] Optionally, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model. The N-BEATS model can be as follows: Figure 4 As shown, the entire model comprises multiple blocks, such as Block1 to Block R. Each block is the most basic structural module of N-BEATS, consisting of multiple fully connected (FC) layers and linear layers.
[0071] In the above embodiments, the Prophet model can fully extract the linear factors in network traffic data, and the N-BEATS model is used to predict the residual of the Prophet model, that is, to obtain the nonlinear factors in network traffic. Then, the prediction results of the two models are added together to improve the accuracy of traffic prediction.
[0072] In the embodiments of this application, please refer to Figure 5 The flowchart illustrates a traffic prediction method provided in an embodiment of this application, which includes the following steps:
[0073] Step 501: Calculate the correlation between the network traffic in the first region and the network traffic in other candidate regions.
[0074] Step 502: Determine the second region from multiple candidate regions based on the calculated correlation.
[0075] Step 503: Obtain the first network traffic data within the first historical time period.
[0076] Step 504: Input the first network traffic data into the first prediction model to obtain the first predicted traffic data of the first region within the target time period.
[0077] Step 505: Obtain the second network traffic data within the second historical time period.
[0078] Step 506: Input the second network traffic data into the first prediction model to obtain the second predicted traffic data for the first region within the first historical period.
[0079] Step 507: Determine the historical residual based on the difference between the second predicted traffic data and the actual network traffic data of the first region in the first historical period.
[0080] Step 508: Input the historical residuals corresponding to the first historical period into the second prediction model to obtain the prediction residuals corresponding to the target period.
[0081] Step 509: Determine the traffic forecast data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0082] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0083] Based on the same inventive concept, this application also provides a traffic prediction apparatus for implementing the traffic prediction method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more traffic prediction apparatus embodiments provided below can be found in the limitations of the traffic prediction method described above, and will not be repeated here.
[0084] In one embodiment, such as Figure 6 As shown, a flow prediction device 600 is provided, including: an acquisition module 601, a first prediction module 602, a second prediction module 603, and a determination module 604, wherein:
[0085] The acquisition module 601 is used to acquire first network traffic data within a first historical time period. The first historical time period is located before the target time period to be predicted. The first network traffic data includes network traffic data of a first region and network traffic data of a second region. The second region is the region where network traffic affects network traffic in the first region.
[0086] The first prediction module 602 is used to input the first network traffic data into the first prediction model to obtain the first predicted traffic data of the first region within the target time period;
[0087] The second prediction module 603 is used to input the historical residual corresponding to the first historical time period into the second prediction model to obtain the prediction residual corresponding to the target time period. The historical residual is used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical time period and the actual network traffic data of the first region in the first historical time period.
[0088] The determining module 604 is used to determine the traffic forecast data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0089] In one embodiment, the apparatus further includes a computing module for calculating the correlation between network traffic in the first region and network traffic in other multiple candidate regions; and determining a second region from the multiple candidate regions based on the calculated correlation.
[0090] In one embodiment, the calculation module is specifically used to calculate, for each candidate region, the covariance between the network traffic data of the candidate region and the network traffic data of the first region, calculate the first variance of the network traffic data of the candidate region, and calculate the second variance of the network traffic data of the first region, and calculate the correlation between the network traffic of the first region and the network traffic of the candidate region based on the covariance, the first variance and the second variance.
[0091] In one embodiment, the first prediction module 602 is specifically used to input the first network traffic data into the first prediction model, decompose the first network traffic data into trend items and random items through the first prediction model, perform traffic prediction processing on the trend items and random items respectively, and fuse the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0092] In one embodiment, the device further includes a residual determination module for acquiring second network traffic data within a second historical time period, the second historical time period being prior to the first historical time period, the second network traffic data including network traffic data of a first region and network traffic data of a second region; inputting the second network traffic data into a first prediction model to obtain second predicted traffic data of the first region within the first historical time period; and determining historical residuals based on the difference between the second predicted traffic data and the actual network traffic data of the first region within the first historical time period.
[0093] In one embodiment, the determining module 604 is specifically used to correct the first predicted flow data based on the prediction residual to obtain flow prediction data.
[0094] In one embodiment, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
[0095] Each module in the aforementioned flow prediction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.
[0096] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 7As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a traffic prediction method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.
[0097] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0098] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring first network traffic data within a first historical time period, the first historical time period being prior to a target time period to be predicted, the first network traffic data including network traffic data of a first region and network traffic data of a second region, wherein the second region is the region where network traffic affects network traffic in the first region; inputting the first network traffic data into a first prediction model to obtain first predicted traffic data for the first region within the target time period; inputting the historical residual corresponding to the first historical time period into a second prediction model to obtain prediction residual corresponding to the target time period, the historical residual being used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical time period and the actual network traffic data of the first region in the first historical time period; and determining traffic prediction data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0099] In one embodiment, when the processor executes the computer program, it further performs the following steps: calculating the correlation between network traffic in the first region and network traffic in multiple candidate regions; and determining a second region from the multiple candidate regions based on the calculated correlation.
[0100] In one embodiment, when the processor executes the computer program, it further performs the following steps: for each candidate region, calculate the covariance between the network traffic data of the candidate region and the network traffic data of the first region, calculate the first variance of the network traffic data of the candidate region, and calculate the second variance of the network traffic data of the first region, and calculate the correlation between the network traffic of the first region and the network traffic of the candidate region based on the covariance, the first variance and the second variance.
[0101] In one embodiment, when the processor executes the computer program, it further performs the following steps: inputting first network traffic data into a first prediction model, decomposing the first network traffic data into trend items and random items through the first prediction model, performing traffic prediction processing on the trend items and random items respectively, and fusing the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0102] In one embodiment, when the processor executes the computer program, it further performs the following steps: acquiring second network traffic data within a second historical period, the second historical period being prior to the first historical period, the second network traffic data including network traffic data of a first region and network traffic data of a second region; inputting the second network traffic data into a first prediction model to obtain second predicted traffic data of the first region within the first historical period; and determining historical residuals based on the difference between the second predicted traffic data and the actual network traffic data of the first region within the first historical period.
[0103] In one embodiment, when the processor executes the computer program, it further performs the following steps: correcting the first predicted flow data based on the prediction residual to obtain flow prediction data.
[0104] In one embodiment, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
[0105] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program performs the following steps: acquiring first network traffic data within a first historical time period, the first historical time period being prior to a target time period to be predicted, the first network traffic data including network traffic data of a first region and network traffic data of a second region, wherein the second region is the region where network traffic affects network traffic in the first region; inputting the first network traffic data into a first prediction model to obtain first predicted traffic data for the first region within the target time period; inputting the historical residual corresponding to the first historical time period into a second prediction model to obtain prediction residual corresponding to the target time period, the historical residual being used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical time period and the actual network traffic data of the first region in the first historical time period; and determining traffic prediction data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0106] In one embodiment, when the computer program is executed by a processor, it further performs the following steps: calculating the correlation between network traffic in a first region and network traffic in multiple candidate regions; and determining a second region from the multiple candidate regions based on the calculated correlation.
[0107] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: for each candidate region, calculating the covariance between the network traffic data of the candidate region and the network traffic data of the first region, calculating the first variance of the network traffic data of the candidate region, and calculating the second variance of the network traffic data of the first region, and calculating the correlation between the network traffic of the first region and the network traffic of the candidate region based on the covariance, the first variance and the second variance.
[0108] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: inputting first network traffic data into a first prediction model, decomposing the first network traffic data into trend items and random items through the first prediction model, performing traffic prediction processing on the trend items and random items respectively, and fusing the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0109] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring second network traffic data within a second historical period, the second historical period being prior to the first historical period, the second network traffic data including network traffic data of a first region and network traffic data of a second region; inputting the second network traffic data into a first prediction model to obtain second predicted traffic data of the first region within the first historical period; and determining historical residuals based on the difference between the second predicted traffic data and the actual network traffic data of the first region within the first historical period.
[0110] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: correcting the first predicted traffic data based on the prediction residual to obtain traffic prediction data.
[0111] In one embodiment, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
[0112] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps: acquiring first network traffic data within a first historical time period, the first historical time period being prior to a target time period to be predicted, the first network traffic data including network traffic data of a first region and network traffic data of a second region, wherein the second region is the region where network traffic affects network traffic in the first region; inputting the first network traffic data into a first prediction model to obtain first predicted traffic data for the first region within the target time period; inputting the historical residual corresponding to the first historical time period into a second prediction model to obtain prediction residual corresponding to the target time period, the historical residual being used to characterize the difference between the network traffic data of the first region predicted by the first prediction model in the first historical time period and the actual network traffic data of the first region in the first historical time period; and determining traffic prediction data for the first region within the target time period based on the first predicted traffic data and the prediction residual.
[0113] In one embodiment, when the computer program is executed by a processor, it further performs the following steps: calculating the correlation between network traffic in a first region and network traffic in multiple candidate regions; and determining a second region from the multiple candidate regions based on the calculated correlation.
[0114] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: for each candidate region, calculating the covariance between the network traffic data of the candidate region and the network traffic data of the first region, calculating the first variance of the network traffic data of the candidate region, and calculating the second variance of the network traffic data of the first region, and calculating the correlation between the network traffic of the first region and the network traffic of the candidate region based on the covariance, the first variance and the second variance.
[0115] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: inputting first network traffic data into a first prediction model, decomposing the first network traffic data into trend items and random items through the first prediction model, performing traffic prediction processing on the trend items and random items respectively, and fusing the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data.
[0116] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring second network traffic data within a second historical period, the second historical period being prior to the first historical period, the second network traffic data including network traffic data of a first region and network traffic data of a second region; inputting the second network traffic data into a first prediction model to obtain second predicted traffic data of the first region within the first historical period; and determining historical residuals based on the difference between the second predicted traffic data and the actual network traffic data of the first region within the first historical period.
[0117] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: correcting the first predicted traffic data based on the prediction residual to obtain traffic prediction data.
[0118] In one embodiment, the first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
[0119] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.
[0120] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0121] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0122] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A traffic prediction method characterized by, The method includes: Calculate the correlation between network traffic in the first region and network traffic in multiple other candidate regions; Based on the calculated correlation, a second region is determined from the plurality of candidate regions; Obtain first network traffic data within a first historical time period, which is located before the target time period to be predicted. The first network traffic data includes network traffic data of the first region and network traffic data of the second region, wherein the second region is the region where network traffic affects network traffic in the first region. The first network traffic data is input into the first prediction model. The first prediction model decomposes the first network traffic data into trend items and random items, and performs traffic prediction processing on the trend items and the random items respectively. The traffic prediction processing results of the trend items and the traffic prediction processing results of the random items are fused to obtain the first predicted traffic data. The historical residual corresponding to the first historical period is input into the second prediction model to obtain the prediction residual corresponding to the target period. The historical residual is used to characterize the difference between the network traffic data of the first region in the first historical period predicted by the first prediction model and the actual network traffic data of the first region in the first historical period. Based on the first predicted traffic data and the predicted residual, the predicted traffic data for the first region during the target time period is determined.
2. The method of claim 1, wherein, The calculation of the correlation between the network traffic of the first region and the network traffic of other candidate regions includes: For each candidate region, the covariance between the network traffic data of the candidate region and the network traffic data of the first region is calculated, the first variance of the network traffic data of the candidate region is calculated, and the second variance of the network traffic data of the first region is calculated. The correlation between the network traffic of the first region and the network traffic of the candidate region is calculated based on the covariance, the first variance, and the second variance.
3. The method of claim 1, wherein, The step of determining the second region from the plurality of candidate regions based on the calculated correlation includes: Sort the absolute values of the correlation between the first region and the plurality of candidate regions from largest to smallest; The second region is selected from a predetermined number of candidate regions before sorting by absolute correlation value.
4. The method of claim 1, wherein, Before inputting the historical residuals corresponding to the first historical period into the second prediction model, the method further includes: Acquire second network traffic data within a second historical time period, the second historical time period being prior to the first historical time period, the second network traffic data including network traffic data of the first region and network traffic data of the second region; The second network traffic data is input into the first prediction model to obtain the second predicted traffic data of the first region within the first historical period. The historical residual is determined based on the difference between the second predicted traffic data and the actual network traffic data of the first region during the first historical period.
5. The method of claim 1, wherein, The step of determining the traffic prediction data for the first region within the target time period based on the first predicted traffic data and the prediction residual includes: The first predicted flow data is corrected based on the predicted residual to obtain the predicted flow data.
6. The method of claim 1, wherein, The first prediction model is the Prophet model, and the second prediction model is the N-BEATS model.
7. A flow prediction device, characterized by, The device includes: The calculation module is used to calculate the correlation between the network traffic of the first region and the network traffic of other multiple candidate regions; and to determine the second region from the multiple candidate regions based on the calculated correlation. The acquisition module is used to acquire first network traffic data within a first historical time period, the first historical time period being before the target time period to be predicted, the first network traffic data including network traffic data of the first region and network traffic data of the second region, wherein the second region is the region where network traffic affects network traffic in the first region; The first prediction module is used to input the first network traffic data into the first prediction model, decompose the first network traffic data into trend items and random items through the first prediction model, perform traffic prediction processing on the trend items and the random items respectively, and fuse the traffic prediction processing results of the trend items and the traffic prediction processing results of the random items to obtain the first predicted traffic data. The second prediction module is used to input the historical residual corresponding to the first historical period into the second prediction model to obtain the prediction residual corresponding to the target period. The historical residual is used to characterize the difference between the network traffic data of the first region in the first historical period predicted by the first prediction model and the actual network traffic data of the first region in the first historical period. The determination module is used to determine the traffic prediction data of the first region within the target time period based on the first predicted traffic data and the prediction residual.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.