Method and device for generating traffic prediction results based on information propagation model of traffic multiplexing redistribution

Abstract

The application relates to a method and device for generating a traffic prediction result based on traffic multiplexing and traffic redivision, an electronic device and a computer readable medium. The method comprises the following steps: configuring parameters of a plurality of information propagation models to be evaluated; dividing the plurality of information propagation models into a plurality of evaluation levels according to a prediction logic; sequentially performing traffic multiplexing and traffic redivision on acquired traffic according to the evaluation levels and inputting the traffic into the plurality of information propagation models to be evaluated; performing calculation on the plurality of information propagation models to generate a plurality of calculation result values; configuring an information propagation strategy for the traffic according to the plurality of calculation result values and acquiring an information propagation result factor; and calculating quality parameters of the plurality of information propagation models according to the information propagation result factor to generate a prediction result. The application can perform traffic multiplexing and traffic redivision on real-time traffic data without affecting the working effect of an online model, thereby accurately and quickly predicting a model to be put online.

Description

Technical Field

[0001] This application relates to the field of computer information processing, and more specifically, to a method, apparatus, electronic device, and computer-readable medium for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division. Background Technology

[0002] In the field of information dissemination, to acquire more high-quality customers and improve service efficiency and quality, it is necessary to filter target customer groups from hundreds of millions of users using information dissemination models for precise information delivery. Therefore, the quality of the information dissemination model becomes one of the key factors in information dissemination. A high-quality information dissemination model can achieve high feedback results with lower information dissemination costs. The quality of the information dissemination model needs to be evaluated through the feedback results obtained.

[0003] In today's information dissemination landscape, traffic is inherently unrepeatable. For example, when a user browses a webpage or short video, a third-party media outlet sends this traffic request to the information disseminator within milliseconds. The disseminator, in turn, must respond within milliseconds, indicating whether or not to proceed with the dissemination. Once the decision to unfollow the user is made, this type of traffic cannot be re-selected or changed, nor can it be reproduced. However, validating the quality of information dissemination models requires a considerable period. Considering these factors, if every update or launch of an information dissemination model involves using the entire volume of online traffic for performance prediction and evaluation, a poor-quality model could lead to reduced information dissemination efficiency and incalculable losses.

[0004] Therefore, there is a need for a new method, apparatus, electronic device, and computer-readable medium for generating traffic prediction results based on a traffic multiplexing and re-division information propagation model.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] In view of this, this application provides a method, apparatus, electronic device, and computer-readable medium for generating traffic prediction results based on an information propagation model of traffic reuse and re-division. This method can reuse and re-divide real-time traffic data without affecting the calculation results of the online model, thereby accurately and quickly predicting multiple models to be launched.

[0007] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0008] According to one aspect of this application, a method for generating traffic prediction results based on a traffic reuse and redistribution information propagation model is proposed. The method includes: configuring model parameters, prediction logic, and quality parameters for multiple information propagation models to be evaluated; dividing the multiple information propagation models into multiple evaluation levels according to the prediction logic; sequentially performing traffic reuse and redistribution on the acquired traffic according to the evaluation levels and inputting it into the multiple information propagation models to be evaluated; the multiple information propagation models calculating the traffic using their corresponding model parameters to generate multiple calculation result values; configuring an information propagation strategy for the traffic based on the multiple calculation result values ​​and obtaining information propagation result factors; calculating the values ​​of the quality parameters corresponding to the multiple evaluation levels based on the information propagation result factors; and determining the prediction result of the corresponding information propagation model based on the value of the quality parameters corresponding to each evaluation level.

[0009] Optionally, the plurality of information propagation models are divided into multiple evaluation levels based on the prediction logic, including: determining multiple evaluation levels according to the prediction logic; and assigning the plurality of information propagation models to the corresponding evaluation levels.

[0010] Optionally, the acquired traffic is repeatedly and re-divided according to the evaluation level and then input into the multiple information propagation models, including: repeatedly using the acquired traffic to input it into the multiple information propagation models corresponding to the evaluation levels; determining the propagation path corresponding to the multiple evaluation levels; and re-dividing the traffic according to the propagation path to input it into the corresponding information propagation model.

[0011] Optionally, the acquired traffic is multiplexed and sequentially input into information propagation models corresponding to multiple evaluation levels, including: multiplexing the acquired traffic to generate multiple traffic streams; and sequentially inputting the multiple traffic streams into information propagation models corresponding to multiple evaluation levels.

[0012] Optionally, determining the propagation path corresponding to the plurality of evaluation levels includes: extracting basic information corresponding to the traffic; performing hash modulo operations on the basic information corresponding to the plurality of evaluation levels respectively; and determining the propagation path corresponding to the plurality of evaluation levels based on the operation results.

[0013] Optionally, the traffic is re-divided according to the propagation path and input into its corresponding information propagation model, including: in each evaluation level, determining the target information propagation model corresponding to the traffic in this level according to the propagation path; and inputting the traffic into the target information propagation model in each evaluation level to achieve traffic re-division.

[0014] Optionally, the process of determining the traffic configuration information propagation strategy and obtaining information propagation result factors based on the multiple calculation result values ​​includes: determining the traffic configuration information propagation strategy based on the multiple calculation result values; and obtaining information propagation result factors corresponding to the information propagation strategy within a preset time range.

[0015] Optionally, the traffic configuration information propagation strategy is based on the multiple calculation results, including: fusing the multiple calculations to generate a fusion result; and using the interval corresponding to the fusion result as the traffic configuration information propagation strategy.

[0016] Optionally, obtaining information propagation result factors corresponding to the information propagation strategy within a preset time range includes: propagating preset information to the terminal corresponding to the traffic according to the information propagation strategy; tracking subsequent instructions of the terminal based on the preset information; and generating the information propagation result factors through the subsequent instructions of the terminal.

[0017] Optionally, the information propagation result factor is generated by subsequent instructions from the terminal, including: recording traffic reuse data and traffic re-division data of the traffic; recording the information propagation strategy corresponding to the traffic; and generating the information propagation result factor based on the traffic reuse data, traffic re-division data, information propagation strategy, and subsequent instructions corresponding to the traffic.

[0018] According to one aspect of this application, an apparatus for generating traffic prediction results based on a traffic reuse and resegmentation information propagation model is proposed. The apparatus includes: a configuration module for configuring model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated; a hierarchy module for dividing the multiple information propagation models into multiple evaluation levels according to the prediction logic; an input module for sequentially performing traffic reuse and resegmentation on the acquired traffic according to the evaluation levels and inputting it into the multiple information propagation models to be evaluated; a calculation module for the multiple information propagation models to calculate the traffic using their corresponding model parameters, generating multiple calculation result values; a strategy module for configuring an information propagation strategy for the traffic based on the multiple calculation result values ​​and obtaining information propagation result factors; and a prediction module for calculating the values ​​of the quality parameters corresponding to the multiple evaluation levels based on the information propagation result factors; and determining the prediction result of the corresponding information propagation model based on the value of the quality parameters corresponding to each evaluation level.

[0019] Optionally, the input module includes: a multiplexing unit for multiplexing the acquired traffic to sequentially input into information propagation models corresponding to multiple evaluation levels; a path unit for determining the propagation path corresponding to the multiple evaluation levels; and a re-division unit for re-dividing the traffic according to the propagation path to input into its corresponding information propagation model.

[0020] According to one aspect of this application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs; and, when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the method as described above.

[0021] According to one aspect of this application, a computer-readable medium is provided having a computer program stored thereon that, when executed by a processor, implements the method described above.

[0022] The method, apparatus, electronic device, and computer-readable medium for generating traffic prediction results based on a traffic reuse and redistribution information propagation model according to this application configure model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated; divide the multiple information propagation models into multiple evaluation levels according to the prediction logic; sequentially reuse and redistribute the acquired traffic according to the evaluation levels and input it into the multiple information propagation models to be evaluated; the multiple information propagation models calculate the traffic with their corresponding model parameters to generate multiple calculation result values; configure an information propagation strategy for the traffic and obtain information propagation result factors based on the multiple calculation result values; calculate the quality parameters of the multiple information propagation models based on the information propagation result factors to generate prediction results. This method enables the reuse and redistribution of real-time traffic data without affecting the calculation result values ​​of the online models, thereby accurately and quickly predicting multiple models to be deployed.

[0023] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0024] The above and other objects, features, and advantages of this application will become more apparent from the detailed description of exemplary embodiments with reference to the accompanying drawings. The drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0025] Figure 1 This is a flowchart illustrating a method for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to an exemplary embodiment.

[0026] Figure 2 This is a schematic diagram illustrating a method for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to another exemplary embodiment.

[0027] Figure 3This is a flowchart illustrating a method for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to another exemplary embodiment.

[0028] Figure 4 This is a flowchart illustrating a method for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to another exemplary embodiment.

[0029] Figure 5 This is a block diagram illustrating an apparatus for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to another exemplary embodiment.

[0030] Figure 6 This is a block diagram illustrating an electronic device according to an exemplary embodiment. Detailed Implementation

[0031] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that this application will be thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted.

[0032] The applicant, through research, discovered that among existing technologies for evaluating information dissemination models that require online data, canary release is the most commonly used technique. When a new information dissemination model is launched and predictions are needed, the number of servers used for project deployment can be controlled, allowing a small portion of traffic to participate in the evaluation of the information dissemination model without affecting other traffic. However, this method still has some problems. For example, when multiple information dissemination models need to be validated, multiple versions need to be released in canary releases. This firstly leads to multiple code versions existing online simultaneously, which is not friendly to system version control; secondly, since each information dissemination model has a long validation cycle, multiple versions validating for a long time will also affect a significant portion of traffic. If the new information dissemination model does not perform well, this part of the traffic will be wasted, resulting in a decrease in the proportion of promotional information. The specific details are shown in the table below, where Master represents the main online model version currently in use:

[0033] Table 1 Grayscale Validation Examples

[0034]

[0035]

[0036] Assuming a total of 20 million customers can receive information daily, and considering the canary rollout testing of four information dissemination models, with each model consuming 5% of the available traffic, as shown in Table 1, the actual promotional effects of algorithms v1.1 and v2.0 are not ideal. The average daily number of users reached is lower than that of the main algorithm version v1.0 used online. This wasted traffic reduces the average daily number of users reached by 110. If the algorithm testing continues for 30 days, this would result in a decrease of 3,300 users. Therefore, canary rollout can also lead to significant traffic waste and a reduction in the number of users reached.

[0037] In this application, addressing the issues of non-reusable and wasteful traffic leading to a decrease in the number of acquired users, a method for predicting information dissemination models based on traffic reuse and redistribution is proposed. By segmenting different traffic streams, layering prediction logic, and fusing results, the reuse of the same traffic stream is achieved. Each stream can participate in the quality evaluation of multiple information dissemination models without affecting the overall message dissemination results, significantly saving the economic and time costs of information dissemination model prediction. This application solves the problems of non-reusable and wasteful traffic, ensuring that the number of users reached does not decrease significantly when validating information dissemination models.

[0038] The content of this application will be described in detail below with the aid of specific embodiments.

[0039] Figure 1 This is a flowchart illustrating a method for generating traffic prediction results based on a traffic multiplexing and re-division information propagation model, according to an exemplary embodiment. The method 10 for generating traffic prediction results based on the traffic multiplexing and re-division information propagation model includes at least steps S102 to S112.

[0040] like Figure 1 As shown, in S102, the model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated are configured. Real-time traffic data from the actual online production environment can be used to simultaneously evaluate multiple information propagation models. Before evaluation, the model parameters of each information propagation model need to be set, as well as the hierarchical relationship between multiple information propagation models (i.e., the prediction logic). Quality parameters used in evaluating the models to be evaluated can also be set, such as information propagation range, information propagation ratio, and propagation feedback effect.

[0041] The prediction logic is used to evaluate the area of ​​traffic input during the experiment. More specifically, in a practical application, different models can also be set with different traffic thresholds.

[0042] In S104, the multiple information propagation models are divided into multiple evaluation levels based on the prediction logic. Multiple evaluation levels can be determined according to the prediction logic; the multiple information propagation models are then assigned to the corresponding evaluation levels.

[0043] In one embodiment, the prediction logic can be configured as shown in Table 2. The first information propagation model will be referred to as Model 1 in the following text and Table 2, the second information propagation model will be referred to as Model 2, and the other models will be configured similarly.

[0044] Table 2 Prediction Logic

[0045]

[0046] As shown in Table 2, in the first-level prediction logic, the currently functioning online Model 1 can be retained. In the second level, two Models 2 can be set, which will be assigned the first traffic threshold and the second traffic threshold respectively. The third level contains three models and their corresponding thresholds. Of course, the information propagation models can also be combined and layered, for example, Model 1, Model 2, and Model 3 can be combined to generate a logical layer.

[0047] In S106, the acquired traffic is repeatedly processed and re-divided according to the evaluation level, and then input into the multiple information propagation models. The acquired traffic can be repeatedly processed and input into the information propagation models corresponding to multiple evaluation levels; the propagation paths corresponding to the multiple evaluation levels are determined; and the traffic is re-divided according to the propagation paths and input into its corresponding information propagation model.

[0048] In the following sections of this application, an actual information dissemination scenario will be used as an example to illustrate the implementation of the technical content of this application. It is understood that the implementation of the technical content of this application can also be applied to other information dissemination scenarios.

[0049] In one embodiment, an information dissemination platform obtains real-time traffic data from a third-party media outlet, then analyzes this traffic data to determine whether to disseminate a preset message. Finally, the result is returned to the third-party media platform, which then disseminates the information.

[0050] More specifically, such as Figure 2 As shown, during the evaluation, third-party media or traffic providers send real-time traffic requests to the information dissemination platform's RTA (Real-Time API) system. The RTA system performs a series of logical checks, such as parameter validation, whitelist verification, and blacklist exclusion. Traffic that remains valid after passing these checks is then sent to the AB-TEST (the core feature of this application's technology) system for experimental prediction.

[0051] Before conducting the evaluation, the algorithm team can configure the prediction logic and traffic ratio allocation. When each piece of traffic passes through the AB-TEST system, the corresponding evaluation log is recorded.

[0052] The content of "reusing and redistributing the acquired traffic according to the evaluation level and inputting it into the multiple information propagation models" will be included in the following sections. Figure 3 The corresponding embodiments are described in detail.

[0053] In S108, the multiple information propagation models calculate the traffic using their corresponding model parameters to generate multiple calculation result values.

[0054] When each traffic item passes through the AB-TEST system, the AB-TEST system will perform traffic reuse and traffic redistribution according to the evaluation level and input the acquired traffic into the multiple information propagation models. The model calculation results will be reported to the corresponding evaluation log for the algorithm team to conduct information propagation model quality analysis.

[0055] In S110, the traffic configuration information propagation strategy is determined based on the multiple calculation results, and an information propagation result factor is obtained. The traffic configuration information propagation strategy can be determined based on the multiple calculation results; an information propagation result factor corresponding to the information propagation strategy within a preset time range is obtained. Finally, the experimental results are fused through the AB-TEST system to determine whether propagation is appropriate and return the result to a third-party media or channel provider.

[0056] The content "configures the information propagation strategy for the traffic based on the multiple calculation result values ​​and obtains the information propagation result factor" will be included in... Figure 4 The corresponding embodiments are described in detail.

[0057] In S112, the values ​​of the quality parameters corresponding to the multiple evaluation levels are calculated based on the information dissemination result factors; and the prediction results of the corresponding information dissemination model are determined based on the prediction results corresponding to each evaluation level.

[0058] In a specific embodiment, when the quality parameter is the information propagation ratio, the information propagation ratio of the prediction logic at each level can be compared to obtain its information propagation effect. Based on the information propagation effect, the propagation information at each level is evaluated first, and then the different information propagation information within each level is evaluated.

[0059] The method for generating traffic prediction results based on the information propagation model of traffic reuse and redistribution according to this application involves configuring model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated; dividing the multiple information propagation models into multiple evaluation levels according to the prediction logic; sequentially reusing and redistributing the acquired traffic according to the evaluation levels and inputting it into the multiple information propagation models to be evaluated; calculating the traffic using the corresponding model parameters of the multiple information propagation models to generate multiple calculation result values; configuring information propagation strategies for the traffic and obtaining information propagation result factors based on the multiple calculation result values; and calculating the quality parameters of the multiple information propagation models based on the information propagation result factors to generate prediction results. This method enables the reuse and redistribution of real-time traffic data without affecting the online model promotion effect, thereby accurately and quickly predicting multiple models to be launched.

[0060] It should be clearly understood that this application describes how specific examples are formed and used, but the principles of this application are not limited to any details of these examples. Rather, based on the teachings of the disclosure of this application, these principles can be applied to many other embodiments.

[0061] Figure 3 This is a flowchart illustrating a method for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to another exemplary embodiment. Figure 3 The process shown in step 30 is... Figure 1 The process shown in the diagram describes in detail S106 "According to the evaluation level, the acquired traffic is reused and re-divided in sequence and input into the multiple information propagation models".

[0062] More specifically, in a practical application, different models can be set with different traffic thresholds, and the prediction logic can be set as shown in Table 3.

[0063] Table 3. Examples of AB-TEST Prediction Logic

[0064]

[0065] like Figure 3 As shown, in S302, the acquired traffic is multiplexed and sequentially input into information propagation models corresponding to multiple evaluation levels. The acquired traffic can be multiplexed to generate multiple traffic streams; these multiple traffic streams are then sequentially input into information propagation models corresponding to multiple evaluation levels.

[0066] A traffic allocation module can be set up within the AB-TEST system of this application. This module allocates traffic to different information propagation models in various logical layers based on the test configuration information when traffic enters. The traffic allocation module is equivalent to a router. All requests entering the traffic allocation module are allocated according to the traffic ID or parameters using the system's randomization algorithm according to a given ratio. The AB-TEST system of this application supports judging a single traffic flow through multiple information propagation models, and also supports multi-threshold experiments for each information propagation model.

[0067] More specifically, in the logical layering, each layer represents a set of predictive logic, which can be used to experiment with an information propagation model at different thresholds. For example, Model 2 is used to evaluate the information propagation model of user quality. After a traffic item passes through this model, a normalized score will be obtained. The experimental group of this layer can be configured with multiple thresholds to conduct experiments and obtain the propagation result of this traffic item through this layer.

[0068] In S304, the propagation paths corresponding to the multiple evaluation levels are determined. Basic information corresponding to the traffic can be extracted; hash modulo operations are performed on the basic information corresponding to each of the multiple evaluation levels; and the propagation paths corresponding to the multiple evaluation levels are determined based on the operation results.

[0069] More specifically, before a traffic item passes through each layer, a hash modulo operation is performed to determine which information propagation model the traffic item should enter in that layer, i.e., which threshold or non-experimental group it should pass through in that layer.

[0070] In S306, the traffic is re-divided according to the propagation path and input into its corresponding information propagation model. At each evaluation level, the target information propagation model corresponding to the traffic at that level is determined based on the propagation path; at each evaluation level, the traffic is input into the target information propagation model to achieve traffic re-division.

[0071] Each traffic segmentation method operates independently. The traffic segmentation algorithm used in each layer hashes the user device number and Layer_id together and takes the modulo to obtain the information propagation model for that traffic allocation at that layer. This avoids the same traffic being input into the same information propagation model at each layer, thus achieving traffic reuse and redistribution.

[0072] The method for generating traffic prediction results based on the information propagation model of traffic reuse and re-division in this application realizes traffic reuse and re-division, multi-level evaluation, and result fusion. Through these methods, traffic reuse is achieved, reducing the economic and time costs of information propagation model prediction.

[0073] Figure 4This is a flowchart illustrating a method for generating traffic prediction results based on an information propagation model for traffic multiplexing and re-division, according to another exemplary embodiment. Figure 4 The process shown in step 40 is... Figure 1 The process shown in the diagram is described in detail in step S110, "Configure the information propagation strategy for the traffic based on the multiple calculation result values ​​and obtain the information propagation result factor".

[0074] like Figure 4 As shown, in S402, the traffic configuration information propagation strategy is determined based on the multiple calculation results. The multiple calculations can be fused to generate a fused result; the traffic configuration information propagation strategy is then determined based on the interval corresponding to the fused result.

[0075] More specifically, the experimental results at each layer are orthogonal and do not affect each other. Traffic will receive a propagation result at each layer. After all layers of experiments are completed, all evaluation results will be merged to obtain the final propagation result, which will then be returned to the media or third-party traffic providers.

[0076] Among them, the fusion methods include commonly used union and intersection operations. The union operation means that as long as there is an experimental result that is propagated, the fusion result is also propagated; the intersection operation indicates that all experimental results are propagated, and only the fusion result is propagated.

[0077] In S404, preset information is propagated to the terminal corresponding to the traffic according to the information propagation strategy.

[0078] In S406, the subsequent instructions of the terminal are tracked based on the preset information.

[0079] In S408, the information propagation result factor is generated through subsequent instructions from the terminal. For example, this may involve recording traffic multiplexing data and traffic re-division data; recording the information propagation strategy corresponding to the traffic; and generating the information propagation result factor based on the traffic multiplexing data, traffic re-division data, information propagation strategy, and subsequent instructions.

[0080] When the AB-TEST system reports logs, it records all predicted partitions that the traffic passes through and the results, providing them to the algorithm team to analyze the merits of all information propagation models.

[0081] Assuming that the prediction logic of Layer 1 and Layer 2 in Table 3 exists in the current AB-TEST system, after a period of experimentation, the merits of multiple information propagation models can be analyzed through data statistics. For example, the evaluation results shown in Table 4 can be obtained through the above embodiment.

[0082] Table 4 Evaluation Results

[0083]

[0084]

[0085] As shown in the table above, information dissemination models 1 and 2 can be compared using groups 1, 2, 3, 4, 5, and 6 respectively. It can be observed that using model 2 (threshold 1) increases the customer acquisition rate by approximately 0.03% compared to not using model 2, while using model 2 (threshold 2) increases the rate by approximately 0.06%. Therefore, with a sufficient sample size, the model ranking can be: Model 2 (using threshold 2) > Model 2 (using threshold 1) > Not using model 2. However, comparing groups 1 and 4, groups 2 and 5, and groups 3 and 6 respectively, it can be found that model 1 does not show a significant advantage over the online model in terms of information dissemination effect. Therefore, when adjusting the information dissemination model used online, both the online model and model 2 (using threshold 2) can be used to jointly determine the traffic dissemination result, thereby increasing the promotion rate of the updated model and achieving a higher dissemination effect.

[0086] The method for generating traffic prediction results based on the information propagation model of traffic reuse and redistribution proposed in this application solves the problems of non-reusable actual traffic, traffic waste, and long verification cycle of information propagation models in the current information propagation and information propagation scenarios by adopting traffic segmentation, multi-layer experiments, and result fusion. It can simultaneously conduct experimental predictions on multiple information propagation models, select high-quality information propagation models, reduce the economic and time costs of information propagation model prediction, and is more conducive to improving the information propagation ratio and rate of return.

[0087] Those skilled in the art will understand that all or part of the steps of the above embodiments are implemented as a computer program executed by a CPU. When the computer program is executed by the CPU, it performs the functions defined by the method provided in this application. The program can be stored in a computer-readable storage medium, such as a read-only memory, a magnetic disk, or an optical disk.

[0088] Furthermore, it should be noted that the above figures are merely illustrative representations of the processes included in the method according to exemplary embodiments of this application, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0089] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0090] Figure 5 This is a block diagram illustrating an apparatus for generating traffic prediction results based on an information propagation model of traffic multiplexing and re-division, according to another exemplary embodiment. Figure 5 As shown, the device 50 for generating traffic prediction results based on the information propagation model of traffic reuse and redistribution includes: a configuration module 502, a hierarchy module 504, an input module 506, a calculation module 508, a strategy module 510, and a prediction module 512.

[0091] The configuration module 502 is used to configure the model parameters, prediction logic, and quality parameters of the multiple information propagation models to be evaluated;

[0092] The hierarchy module 504 is used to divide the multiple information propagation models into multiple evaluation levels according to the prediction logic; the hierarchy module 504 is also used to determine multiple evaluation levels according to the prediction logic; and to divide the multiple information propagation models into the corresponding evaluation levels.

[0093] The input module 506 is used to perform traffic reuse and traffic re-division on the acquired traffic in sequence according to the evaluation level and input it into multiple information propagation models to be evaluated;

[0094] The input module 506 includes:

[0095] The multiplexing unit is used to multiplex the acquired traffic so that it can be sequentially input into the information propagation model corresponding to multiple evaluation levels.

[0096] A path unit is used to determine the propagation path corresponding to the multiple evaluation levels;

[0097] The re-splitting unit is used to re-split the traffic according to the propagation path so as to input its corresponding information propagation model.

[0098] The calculation module 508 is used by the multiple information propagation models to calculate the traffic with their corresponding model parameters and generate multiple calculation result values;

[0099] The strategy module 510 is used to configure the traffic information propagation strategy based on the multiple calculation results and obtain the information propagation result factor; the strategy module 510 is also used to configure the traffic information propagation strategy based on the multiple calculation results and obtain the information propagation result factor corresponding to the information propagation strategy within a preset time range.

[0100] The prediction module 512 is used to calculate the prediction results corresponding to the multiple evaluation levels based on the information dissemination result factors; and to determine the prediction results of the corresponding information dissemination model based on the prediction results corresponding to each evaluation level.

[0101] The apparatus for generating traffic prediction results based on the information propagation model of traffic reuse and redistribution according to this application configures the model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated; divides the multiple information propagation models into multiple evaluation levels according to the prediction logic; sequentially performs traffic reuse and redistribution on the acquired traffic according to the evaluation levels and inputs it into the multiple information propagation models to be evaluated; the multiple information propagation models calculate the traffic with their corresponding model parameters to generate multiple calculation result values; configures an information propagation strategy for the traffic and obtains information propagation result factors based on the multiple calculation result values; calculates the quality parameters of the multiple information propagation models based on the information propagation result factors, thereby generating prediction results. This method enables the reuse and redistribution of real-time traffic data without affecting the calculation result values ​​of the online models, thus accurately and quickly predicting multiple models to be launched.

[0102] like Figure 6 As shown, this application provides an electronic device including a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640.

[0103] Memory 630 is used to store computer programs;

[0104] When the processor 610 executes the program stored in the memory 630, it implements the method for generating traffic prediction results based on the information propagation model of traffic multiplexing and re-division in any of the above embodiments.

[0105] The communication interface 620 is used for communication between the above-mentioned electronic devices and other devices.

[0106] The memory 630 may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory 630 may also be at least one storage device located remotely from the aforementioned processor 610.

[0107] This application provides a computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement a method for generating traffic prediction results based on a traffic multiplexing and resegmentation information propagation model according to any of the above embodiments. For example, the method involves configuring model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated; dividing the multiple information propagation models into multiple evaluation levels according to the prediction logic; sequentially performing traffic multiplexing and resegmentation on the acquired traffic according to the evaluation levels and inputting it into the multiple information propagation models to be evaluated; the multiple information propagation models calculating the traffic using their corresponding model parameters to generate multiple calculation result values; configuring an information propagation strategy for the traffic and obtaining information propagation result factors based on the multiple calculation result values; calculating the values ​​of the quality parameters corresponding to the multiple evaluation levels based on the information propagation result factors; and determining the prediction result of the corresponding information propagation model based on the value of the quality parameters corresponding to each evaluation level.

[0108] Exemplary embodiments of this application have been specifically shown and described above. It should be understood that this application is not limited to the detailed structures, arrangements, or implementation methods described herein; rather, this application is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended claims.

Claims

1. A method for generating traffic prediction results based on an information propagation model for traffic reuse and redistribution, characterized in that, include: Configure the model parameters, prediction logic, and quality parameters for the multiple information propagation models to be evaluated; Based on the prediction logic, the multiple information propagation models are divided into multiple evaluation levels; The acquired traffic is reused and re-divided according to the evaluation level, and then input into the multiple information propagation models. The multiple information propagation models calculate the traffic using their corresponding model parameters, generating multiple calculation result values; Multiple calculation results are fused to generate a fusion result. The fusion methods include union and intersection operations. Union means that as long as there is an experimental result that propagates, the fusion result is also propagated. Intersection means that if all experimental results propagate, then the fusion result is propagated. The traffic configuration information propagation strategy is based on the interval corresponding to the fusion result. According to the information propagation strategy, preset information is propagated to the terminal corresponding to the traffic. The terminal's subsequent instructions are tracked based on the preset information; Record the traffic reuse data and traffic re-division data of the traffic; Record the information dissemination strategy corresponding to the traffic; Based on the traffic reuse data, traffic re-division data, information propagation strategy, and subsequent instructions corresponding to the traffic, an information propagation result factor is generated. Calculate the values ​​of the quality parameters corresponding to the multiple evaluation levels based on the information dissemination result factors; The prediction results of the corresponding information dissemination model are determined based on the values ​​of the quality parameters corresponding to each assessment level.

2. The method as described in claim 1, characterized in that, Based on the prediction logic, the multiple information propagation models are divided into multiple evaluation levels, including: Multiple evaluation levels are determined according to the prediction logic; The multiple information dissemination models are divided into corresponding evaluation levels.

3. The method as described in claim 1, characterized in that, The acquired traffic is then reused and re-divided according to the evaluation level, and input into the multiple information propagation models, including: The acquired traffic is reused and then sequentially input into the information propagation models corresponding to multiple evaluation levels. Determine the propagation paths corresponding to the multiple evaluation levels; The traffic is re-divided according to the propagation path and input into its corresponding information propagation model.

4. The method as described in claim 3, characterized in that, The acquired traffic is multiplexed and sequentially input into information propagation models corresponding to multiple evaluation levels, including: The acquired traffic is reused to generate multiple traffic streams; The multiple traffic streams are sequentially input into the information propagation models corresponding to the multiple evaluation levels.

5. The method as described in claim 3, characterized in that, Determining the propagation paths corresponding to the multiple evaluation levels includes: Extract the basic information corresponding to the traffic; The basic information is hashed and moduloed for each of the multiple evaluation levels. The propagation path of the flow is determined based on the calculation results at the multiple evaluation levels.

6. The method as described in claim 3, characterized in that, The traffic is re-divided according to the propagation path to input its corresponding information propagation model, including: In each evaluation level, the target information propagation model corresponding to the traffic in this level is determined based on the propagation path; The traffic is input into the target information propagation model at each evaluation level to achieve traffic re-division.

7. An apparatus for generating traffic prediction results based on a traffic reuse and redistribution information propagation model, characterized in that, include: The configuration module is used to configure the model parameters, prediction logic, and quality parameters of multiple information propagation models to be evaluated. A hierarchy module is used to divide the multiple information propagation models into multiple evaluation levels based on the prediction logic; The input module is used to perform traffic reuse and traffic redistribution on the acquired traffic according to the evaluation level and input it into multiple information propagation models to be evaluated. The calculation module is used to calculate the traffic using the multiple information propagation models with their corresponding model parameters, and generate multiple calculation result values; The strategy module is used to fuse multiple calculated values ​​to generate a fusion result. The fusion methods include union and intersection operations. Union operation means that if any experimental result is propagated, the fusion result is propagated; intersection operation means that if all experimental results are propagated, the fusion result is propagated. The module configures an information propagation strategy for the traffic based on the interval corresponding to the fusion result; propagates preset information to the terminal corresponding to the traffic according to the information propagation strategy; tracks subsequent instructions from the terminal based on the preset information; records traffic reuse data and traffic re-division data of the traffic; records the information propagation strategy corresponding to the traffic; and generates an information propagation result factor based on the traffic reuse data, traffic re-division data, information propagation strategy, and subsequent instructions corresponding to the traffic. The prediction module is used to calculate the values ​​of the quality parameters corresponding to the multiple evaluation levels based on the information dissemination result factors; and to determine the prediction result of the corresponding information dissemination model based on the value of the quality parameter corresponding to each evaluation level.

8. The apparatus as claimed in claim 7, characterized in that, The input module includes: The multiplexing unit is used to multiplex the acquired traffic so that it can be sequentially input into the information propagation model corresponding to multiple evaluation levels. A path unit is used to determine the propagation path corresponding to the multiple evaluation levels; The re-splitting unit is used to re-split the traffic according to the propagation path so as to input its corresponding information propagation model.

9. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 6.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 6.

Images