Method, device and equipment for determining traffic use behavior of user and storage medium

Abstract

The invention discloses a method, device and equipment for determining a user traffic use behavior and a storage medium, and relates to the technical field of communication. The method comprises the following steps: acquiring data of a flow use behavior of a user in a current time period and a historical behavior change probability; respectively calculating the distance between the data of the traffic use behavior and the data of a plurality of preset traffic use behaviors, and determining the use behavior probability that the data of the traffic use behavior belongs to each preset traffic use behavior according to the relationship between the distance and the user use behavior probability; calculating the user target use behavior probability of the user in the next time period according to the user use behavior probability and the historical behavior change probability; and determining the traffic use behavior of the user in the next time period according to the user target use behavior probability of the user in the next time period of the current time period. According to the embodiment of the invention, the traffic use mode is captured by combining the use behavior probability and the historical behavior change probability, so that the accuracy of determining the traffic use behavior is improved.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to methods, devices, equipment, and computer storage media for determining user traffic usage behavior. Background Art

[0002] The widespread adoption and development of 5G networks presents numerous challenges, particularly in determining user data usage. Due to widespread WiFi coverage, 5G data usage exhibits high levels of discontinuity and multiple breakpoints, making it more difficult to determine user behavior.

[0003] When determining users' traffic usage behavior, existing technologies mostly rely on statistical learning or simple machine learning models. This method performs well in environments with large and stable data volumes, but is less effective when dealing with rapidly changing user usage behaviors. Due to the great uncertainty and instability of users' 5G traffic usage, existing technologies find it difficult to accurately capture users' irregular usage patterns, resulting in large errors between the determination results and the actual situation, and low determination accuracy. Summary of the Invention

[0004] The embodiments of the present application provide a method, apparatus, device, and computer storage medium for determining user traffic usage behavior to solve the problem of low accuracy of traffic usage determination models in the prior art.

[0005] In a first aspect, an embodiment of the present application provides a method for determining user traffic usage behavior, the method comprising:

[0006] Obtain the user's traffic usage behavior data for the current time period and the historical behavior change probability. The historical behavior change probability is the probability that the user's traffic usage behavior will shift from one period in the historical time period to the next.

[0007] Calculate the distance between the data of the traffic usage behavior and the data of multiple preset traffic usage behaviors respectively, and determine the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distance and the probability of the user's usage behavior;

[0008] Calculate the user's target usage behavior probability in the next time period of the current time period based on the user's usage behavior probability and the historical behavior change probability;

[0009] The user's traffic usage behavior in the next time period is determined based on the user's target usage behavior probability in the next time period of the current time period.

[0010] In a second aspect, an embodiment of the present application provides a device for determining user traffic usage behavior, the device comprising:

[0011] An acquisition module is used to obtain the user's traffic usage behavior data for the current time period and the historical behavior change probability. The historical behavior change probability is the probability that the user's traffic usage behavior will shift from one period in the historical time period to the next period.

[0012] a determination module, configured to respectively calculate the distance between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors, and determine the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distance and the probability of the user's usage behavior;

[0013] A calculation module is used to calculate the probability of the user's target usage behavior in the next time period of the current time period based on the user's usage behavior probability and the historical behavior change probability;

[0014] The determination module is used to determine the user's traffic usage behavior in the next time period based on the user's target usage behavior probability in the next time period of the current time period.

[0015] In a third aspect, an embodiment of the present application provides a terminal device, comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the method for determining user traffic usage behavior as in the first aspect is implemented.

[0016] In a fourth aspect, an embodiment of the present application provides a computer storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the method for determining user traffic usage behavior as in the first aspect is implemented.

[0017] In a fifth aspect, an embodiment of the present application provides a computer program product. When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the method for determining user traffic usage behavior as in the first aspect.

[0018] The method for determining user traffic usage behavior provided by the embodiment of the present application first obtains the user's traffic usage behavior data for the current time period and the historical behavior change probability, where the historical behavior change probability is the probability that the user's traffic usage behavior shifts from one period in the historical time period to the next; then calculates the distance between the traffic usage behavior data and the data of multiple preset traffic usage behaviors, and determines the probability that the traffic usage behavior data belongs to each preset traffic usage behavior based on the relationship between the distance and the user's usage behavior probability; calculates the user's target usage behavior probability for the next time period based on the user's usage behavior probability and the historical behavior change probability; and determines the user's traffic usage behavior for the next time period based on the user's target usage behavior probability for the next time period. By obtaining traffic usage behavior data, a comprehensive understanding of user behavior patterns can be achieved. The behavior change probability reflects the user's behavior shift trend in different time periods and can capture the dynamic changes in user behavior. By calculating the distance and converting the distance into user behavior probability, the likelihood of the user falling into different behavior patterns can be evaluated in the form of probabilities. Combining the behavior probability and the historical behavior change probability, the user's behavior probability and corresponding usage behavior for the next time period can be dynamically predicted, thereby capturing the user's traffic usage pattern and improving the accuracy of traffic usage behavior determination. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 This is a flow chart of a method for determining user traffic usage behavior provided by an embodiment of the present application;

[0021] Figure 2 This is a flow chart of an implementation method for determining the probability of a user's target usage behavior provided by an embodiment of the present application;

[0022] Figure 3 This is a flow chart of an implementation method of flow determination model training provided in an embodiment of the present application;

[0023] Figure 4 Schematic diagram of the structure of the device for determining user traffic usage behavior provided by an embodiment of the present application;

[0024] Figure 5 It is a structural diagram of the terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present application by illustrating the examples of the present application.

[0026] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0027] Existing traffic prediction models, whether continuous or intermittent, often struggle to accurately capture these irregular usage patterns, and are even less able to effectively explain and predict user behavior. Faced with this situation, existing technical solutions often rely on statistical learning or simple machine learning models. While these methods perform well in environments with large amounts of data and stability, they struggle to cope with rapidly changing user behavior. For example, these models often fail to effectively distinguish whether a user's reduced 5G usage is due to a change in preferences or simply because they have chosen alternative services such as WiFi in the short term. Furthermore, traffic determination models in existing methods often focus on the overall identification and prediction of 5G intelligent network traffic, prioritizing global analysis from a business perspective. They fail to delve into the details of individual users' traffic usage, making it impossible to accurately predict and reasonably explain individual users, resulting in low accuracy in traffic behavior determination.

[0028] To address the problems of the prior art, embodiments of the present application provide a method for determining user traffic usage behavior. The method first obtains the user's traffic usage behavior data for the current time period and the historical behavior change probability, where the historical behavior change probability is the probability that the user's traffic usage behavior shifts from one period in the historical time period to the next. The method then calculates the distance between the traffic usage behavior data and multiple preset traffic usage behavior data, and determines the probability that the traffic usage behavior data belongs to each preset traffic usage behavior based on the relationship between the distance and the user's usage behavior probability. The method then calculates the user's target usage behavior probability for the next time period based on the user's usage behavior probability and the historical behavior change probability. The method then determines the traffic usage behavior corresponding to the user's target usage behavior probability that is not less than a set threshold as the user's traffic usage behavior for the next time period. By obtaining traffic usage behavior data, a comprehensive understanding of user behavior patterns can be achieved. The behavior change probability reflects the user's behavior shift trends over different time periods, allowing the user to capture dynamic changes in behavior. By calculating the distance and converting the distance into a user behavior probability, the likelihood of the user falling into different behavior patterns can be evaluated in a probabilistic manner. Combining the behavior probability and the historical behavior change probability, the user's behavior probability and corresponding usage behavior for the next time period can be dynamically predicted, thereby capturing the user's traffic usage pattern and improving the accuracy of traffic usage behavior determination.

[0029] The following first introduces the method for determining user traffic usage behavior provided by an embodiment of the present application.

[0030] Figure 1 FIG2 shows a flow chart of a method for determining user traffic usage behavior provided by an embodiment of the present application. Figure 1 As shown, the method may include the following steps: S101 to S104.

[0031] S101, obtaining the user's traffic usage behavior data for the current time period and the historical behavior change probability, where the historical behavior change probability is the probability that the user's traffic usage behavior shifts from one period in the historical time period to the next period.

[0032] Among them, the data on traffic usage behavior in the current time period is the data on the user's traffic usage behavior within the recent specified time, such as this week or this month; the data on traffic usage behavior may include user information data and traffic usage data, user information data may include user basic feature information such as user age and user gender, and traffic usage data may include traffic usage amount and traffic usage frequency, etc.; the probability of historical behavior change is the probability of transfer of the user's historical behavior model.

[0033] In some embodiments, the historical behavior change probability can be a Markov transition probability, which is generally estimated based on the frequency of historical data. For example, the historical data includes the number of people in group k who used traffic at time t, and the number of people in group k who switched to group m at time t+1. The ratio of the two can be calculated to obtain the probability that group k will switch to group m at time t+1.

[0034] In some embodiments, the historical behavior change probability can be obtained through historical data statistics. The historical behavior change probability can be the probability of a period in any historical time period shifting to the next period. For example, if the time period is one month and the current time period is August, the historical behavior change probability can be selected as the probability of shifting from April to May, or the probability of shifting from June to July.

[0035] The probability of historical behavior changes supports behavior tracking across time periods and can be applied to long-term user behavior analysis related to seasonal fluctuations, etc.

[0036] In one example, traffic usage behavior includes using traffic and not using traffic, and the historical behavior change probability is the probability that the user changes from using traffic to using traffic or not using traffic, and the user changes from not using traffic to not using traffic or using traffic.

[0037] By obtaining the probability of historical behavior changes, we can capture the possibility of changes in user behavior and provide a mathematical basis for subsequent predictions.

[0038] S102, respectively calculating the distance between the data of the traffic usage behavior and the data of multiple preset traffic usage behaviors, and determining the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distance and the probability of the user's usage behavior;

[0039] The preset traffic usage behavior is a predefined category of typical user behavior, such as 5G traffic users, non-5G traffic users, heavy nighttime users, and heavy daytime users. The distance between traffic usage data and the preset traffic usage data is an indicator measuring the similarity between the user's traffic usage behavior data and the preset traffic usage behavior.

[0040] In some embodiments, the preset traffic usage behavior data is historical traffic usage behavior data of a cluster center obtained by clustering data of historical traffic usage behaviors.

[0041] In some embodiments, the distance between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors may be measured using Euclidean distance or cosine similarity.

[0042] By measuring the behavioral similarity between user traffic usage behavior data and preset traffic usage behavior, the bias of manual classification is reduced.

[0043] S103 , calculating the user's target usage behavior probability in the next time period of the current time period based on the user's usage behavior probability and the historical behavior change probability.

[0044] The user target usage behavior probability is the probability that the user belongs to a certain behavior category in the next cycle.

[0045] The user's target usage behavior probability is calculated by the user's usage behavior probability and the historical behavior change probability, combining individual behavior trends and group transfer laws, and comprehensively considering the common behavior and individual behavior of the group, thereby improving the prediction accuracy.

[0046] S104 , determining the traffic usage behavior of the user in the next time period according to the probability of the user's target usage behavior in the next time period of the current time period.

[0047] In some embodiments, the traffic usage behavior corresponding to the user's target usage behavior probability that is not less than a set threshold can be determined as the user's traffic usage behavior in the next time period.

[0048] In some embodiments, the threshold may be set to the highest user target usage behavior probability, that is, the traffic usage behavior corresponding to the highest target user usage behavior probability is selected as the user's traffic usage behavior in the next time period.

[0049] By acquiring data on traffic usage behavior, we can fully understand user behavior patterns. The probability of behavior change reflects the trend of user behavior shifts over different time periods and can capture the dynamic changes in user behavior. By calculating distance and converting it into user behavior probability, we can probabilistically assess the likelihood of users following different behavior patterns. Combining behavior probability with the probability of historical behavior change, we can dynamically predict the user's behavior probability and corresponding usage behavior in the next time period, thereby capturing the user's traffic usage patterns and improving the accuracy of traffic usage behavior determination.

[0050] In some embodiments, obtaining data on the user's traffic usage behavior in the current time period may include:

[0051] Obtain user information data and traffic usage data;

[0052] Extracting field features corresponding to user information data and user traffic usage data in user behavior data based on preset feature fields;

[0053] According to the user identifier, the field features corresponding to the user information data and the user traffic usage data are associated to form a feature set to obtain the user features.

[0054] In some embodiments, user traffic data can be filtered through 5G network usage records and 5G package purchase records. 5G network usage records are derived from sources such as base station logs and user device reported data. By querying relevant fields in the database such as user access network type and service type, users with 5G network access records within a specified time period are filtered out. 5G package purchase records are retrieved from data sources such as billing systems and service subscription records to find users of purchased and effective 5G packages, confirming that the user has subscribed to a 5G exclusive package or a mixed package that includes 5G services, and that the package status is activated or in use.

[0055] In some embodiments, obtaining user information data may include:

[0056] Extract each user's unique identifier (such as user ID, mobile phone number, etc.) from the database to ensure that each user has a unique and identifiable identity in the system;

[0057] By calling the database interface or application programming interface (API), the encrypted variables are associated with the user information table, package service table, broadband registration table, activity participation record table, etc., and the user statistical attribute data corresponding to the encrypted variables are retrieved;

[0058] Extract the user's basic statistical attributes from the associated data, including age, gender, occupation, geographic location, and device type.

[0059] In some embodiments, extracting field features corresponding to user information data in user behavior data based on preset feature fields may include:

[0060] Design the user feature table structure, which can include at least one of the following: encrypted variables in fields, age or age range, gender, occupation or occupation label, geographic location information such as latitude and longitude, city level, administrative division, urban and rural classification, device type, device details, and network access age;

[0061] Clean and organize the queried user statistical attribute data according to the designed table structure, and import them into the newly created user feature table in batches;

[0062] Perform integrity and consistency checks on the imported data to ensure that the data corresponding to each encrypted variable is complete and conflict-free.

[0063] In one example, a user basic information feature table is established as shown in Table 1, which contains key statistical attribute information of the user and provides basic data support for subsequent usage behavior analysis and other links.

[0064] Table 1 User basic information characteristics table

[0065]

[0066] In some embodiments, obtaining user traffic data may include:

[0067] Through the telecom operator's internal system interface or data warehouse, using encrypted variables as indexes, obtain the user's historical daily traffic usage data, including usage timestamp, data volume, usage frequency, usage amount and other information.

[0068] In some embodiments, the unique identifier may be encrypted to generate a new encrypted variable, using asymmetric encryption, hash functions, tokenization, and other encryption methods. Decryption testing is performed on some of the encrypted encrypted variables to ensure that the encryption process was correct and that the original user unique identifier can be restored after decryption. The user's identity cannot be directly identified in the encrypted state.

[0069] In some embodiments, extracting field features corresponding to user traffic usage data based on preset feature fields may include:

[0070] Create a user 5G usage behavior table, including at least one of the basic usage behavior data fields such as encrypted variables, usage timestamp, data volume, usage frequency, and usage amount;

[0071] Fill the collected and organized historical traffic usage daily frequency data into the user's 5G usage behavior table to ensure that the data corresponding to each encrypted variable is complete and correct;

[0072] Fill multi-period customer value analysis and customer relationship management (Recency, Frequency, Monetary, RFM) indicators into the corresponding new fields of the user 5G usage behavior table, and perform data consistency checks and logical verification on the updated table to ensure that the newly added multi-period RFM indicators are calculated correctly and correspond correctly to the encrypted variables in the user basic information table.

[0073] In an example, a user 5G traffic usage behavior feature table is established as shown in Table 2. The table reflects the user's 5G traffic usage habits and changes in the past six months, which helps to predict the user's 5G traffic usage behavior.

[0074] Table 2 User 5G traffic usage behavior characteristics

[0075]

[0076]

[0077] In some embodiments, multi-period customer value analysis and customer relationship management (Recency, Frequency, Monetary, RFM) indicators are populated into corresponding new fields of the user 5G usage behavior table, including:

[0078] Segment historical traffic using daily frequency data according to the time intervals of the most recent week, month, three months, and six months, thereby dividing the time window;

[0079] In each time window, the difference between the last time the user used 5G traffic and the end time of the window is calculated in days to determine the most recent usage time (R);

[0080] Count the total time users spend using 5G services in each time window on an hourly basis to determine the frequency of use (F);

[0081] Summarize the total traffic used by users on 5G services in each time window to determine the traffic used (M);

[0082] The usage amount is determined by converting the current usage M into fees according to the user's current package or traffic billing model.

[0083] In some embodiments, before respectively calculating the distances between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors, the method may further include:

[0084] Fill missing values in the data on traffic usage behavior with zero;

[0085] Identify outliers in traffic usage data and delete, replace, or correct them;

[0086] Standardize the numerical data in the data on traffic usage behavior.

[0087] Filling missing values in traffic usage behavior data with zeros can keep the data set complete, avoid erroneous results due to missing values, and enhance data availability; identifying and processing outliers in traffic usage behavior data can remove data points that do not conform to normal traffic usage patterns, so that the data can more accurately reflect the actual traffic usage and improve the accuracy of subsequent analysis; standardizing numerical data and converting data with different features to the same scale can eliminate the impact of different features due to different dimensions or numerical ranges.

[0088] In some embodiments, when performing outlier identification, identification methods such as the box plot method and the standard score Z-score method can be used.

[0089] Numerical data can be Z-score standardized, and the standardization formula is:

[0090]

[0091] Among them, x ij is the original value of user i on feature j, μ j is the mean of feature j, σ j is the standard deviation of feature j.

[0092] In some embodiments, determining the probability that data on traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between distance and user usage behavior probability may include:

[0093] Calculate the sum of the distances between the data of the traffic usage behavior and the data of multiple preset traffic usage behaviors to obtain an overall distance;

[0094] Calculating the ratio of the distance between the data of the traffic usage behavior and the data of each preset traffic usage behavior to the overall distance, respectively, to obtain a normalized result of the distance between the data of the traffic usage behavior and the data of each preset traffic usage behavior;

[0095] According to the relationship between the normalization result and the probability of user usage behavior, the probability of user usage behavior that the data of traffic usage behavior belongs to the preset behavior corresponding to the data of each preset traffic usage behavior is determined.

[0096] By calculating the overall distance and normalizing each distance, the possible dimensional differences between different distance values are eliminated, and the similarity between user traffic usage behavior and preset behavior is accurately quantified. The user usage behavior probability is determined by the relationship between the normalized result and the user usage behavior probability, and the distance indicator is converted into an intuitive probability value, providing a basis for subsequent decision-making.

[0097] In some embodiments, the calculation formula for the distance between the data of the traffic usage behavior and the data of the preset traffic usage behavior may be:

[0098] Distance i,k =||x i -μ k || 2

[0099] Among them, Distance i,k is the distance between the traffic usage behavior data of the i-th user and the k-th preset traffic usage behavior data, x i is the data of traffic usage behavior of the i-th user, μ k The data of the k-th preset traffic usage behavior.

[0100] In some embodiments, the calculation formula for the user usage behavior probability may be:

[0101]

[0102] in, is the probability that the traffic usage behavior data of the i-th user belongs to the k-th preset traffic usage behavior group, It is the normalized result of the distance between the i-th user traffic usage behavior data and the k-th preset traffic usage behavior data.

[0103] Since the distance value represents the similarity between the user's traffic usage behavior data and the preset traffic usage behavior data, the larger the distance value, the smaller the similarity. Therefore, the larger the normalized result, the smaller the probability that the user's traffic usage behavior data belongs to the corresponding preset traffic usage behavior group. Therefore, by subtracting the normalized result from 1, the probability of positively correlated user usage behavior can be obtained.

[0104] In some embodiments, as Figure 2 As shown, calculating the user's target usage behavior probability in the next time period according to the user's usage behavior probability and the historical behavior change probability may include: S201 and S202.

[0105] S201 : Determine the probability of a user's first traffic usage behavior in the next time period based on the user's usage behavior probability and the historical behavior change probability.

[0106] The first traffic usage behavior data is the traffic behavior usage probability of the next time period obtained by the user at the group level.

[0107] In some embodiments, when the preset traffic usage behavior includes using 5G traffic and not using 5G traffic, the calculation formula for the first traffic usage behavior probability can be:

[0108]

[0109] in, is the probability of user i switching to the first traffic usage behavior of “using 5G traffic” at time t+1, is the probability of user i switching to the first traffic usage behavior of “not using 5G traffic” at time t+1, User i belongs to group group at time t k The probability of is the historical behavior change probability of transferring from the state of group k to the state of “using 5G traffic”, which is, is the historical behavior change probability of transitioning from the state of group k to the “not using 5G traffic” state.

[0110] S202 , determining a user target usage behavior probability of the user in the next time period according to the first traffic usage behavior probability, traffic usage behavior data of the current time period, and a preset weight vector.

[0111] The user target usage behavior probability is the user target usage behavior probability obtained by adjusting the user's first traffic behavior usage probability for the next time period obtained at the group level according to individual circumstances.

[0112] In some embodiments, the formula for calculating the probability of user target usage behavior may be:

[0113]

[0114] in, is the weight vector of the relationship between the user's traffic usage behavior data and the preset traffic usage behavior, β is the weight coefficient of the user's traffic usage behavior data on the traffic usage behavior, γ is the weight coefficient of the group prediction result on the traffic usage behavior, and σ is the sigmoid function, which maps the predicted value to the (0,1) interval, indicating the probability of the user using traffic at time t+1.

[0115] In some embodiments, the coefficients in the user target usage behavior probability calculation formula can be learned through training data, using the cross-validation method, and by maximizing the prediction accuracy evaluation index to learn the fusion coefficients β and γ of the group prediction results and the individual characteristics in the individual usage behavior prediction.

[0116] The common patterns of user groups are obtained through the probability of first traffic usage behavior, and the general patterns of user status changes from a group perspective are obtained. At the same time, combined with the preset weights of users' individual characteristics and the differences in users' individual characteristics, the common behaviors and individual behaviors of the group are comprehensively considered, thereby improving the prediction accuracy.

[0117] In some embodiments, before respectively calculating the distances between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors, and determining the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distances and the user usage behavior probabilities, the method may further include:

[0118] Obtain data on training traffic usage behavior;

[0119] The data of the training traffic usage behavior that meets the preset conditions are selected as the initial cluster center, and based on the distance between the data of the training traffic usage behavior other than the data of the training traffic usage behavior corresponding to the initial cluster center and the data of the initial cluster center, the data of the training traffic usage behavior other than the data of the training traffic usage behavior corresponding to the initial cluster center are assigned to the cluster where the nearest initial cluster center is located, thereby obtaining multiple clusters;

[0120] Calculate the feature mean of all training traffic usage behavior data in each cluster and update the cluster center based on the feature mean;

[0121] When the cluster center no longer changes, the data of the traffic usage behavior corresponding to the cluster center is determined to be the preset traffic usage behavior data.

[0122] Use the median / mode to initialize the cluster center to avoid the local optimal problem caused by random initialization and improve clustering stability; use the clustering algorithm to automatically adjust the cluster center according to the training data to reduce manual intervention, and at the same time adapt to different user behavior patterns based on different training data.

[0123] In some embodiments, the clustering method may be a K-means clustering algorithm.

[0124] In some embodiments, the preset condition may be the median and / or mode. When selecting the median and / or mode training traffic usage behavior data as the initial cluster center, weights may be assigned based on the different data of traffic usage behavior to highlight the importance of different features.

[0125] In some embodiments, after determining the traffic usage behavior corresponding to the user's target usage behavior probability that is not less than a set threshold as the user's traffic usage behavior in the next time period, the method may further include:

[0126] The data of the user's traffic usage behavior in the current time period is input into the traffic determination model, and the traffic determination model is used to determine the traffic usage value in the next time period based on the relationship between the data of the traffic usage behavior in the current time period, the pre-trained parameters of the traffic determination model and the traffic usage value.

[0127] The user's traffic usage behavior data is input into the traffic determination model. The model can predict future traffic usage based on the user's historical behavior patterns and characteristics, thereby realizing traffic usage prediction.

[0128] In some embodiments, the traffic determination model may be a gradient boosting tree model, which is optimized by gradually adding models with poor prediction performance.

[0129] In some embodiments, after determining the traffic usage value based on the data of traffic usage behavior, the pre-trained parameters of the traffic determination model, and the relationship between the traffic usage value, the method may further include:

[0130] According to the traffic usage value, the clustering results and the probability of historical behavior changes are reversely adjusted. The adjustment formula is:

[0131]

[0132] Among them, mape i is the MAPE value of the i-th sample, which measures the predicted value and the true value y i The difference between is the traffic prediction value, y i is the flow usage value, is the probability of user i switching to the first traffic usage behavior of “using 5G traffic” at time t+1, is the probability of user i switching to the first traffic usage behavior of “not using 5G traffic” at time t+1.

[0133]

[0134] in, is the average MAPE value of all samples, and N is the total number of samples.

[0135]

[0136] Among them, AIC is a statistic used to weigh the goodness of fit and complexity of the model in model selection, X i is the feature vector of the i-th sample, C i is the cluster center of the i-th sample, k is the number of cluster centers, d is the dimension of the feature, and n is the number of samples.

[0137] In some embodiments, as Figure 3 As shown, before determining the traffic usage behavior corresponding to the user target usage behavior probability not less than a set threshold as the user's traffic usage behavior in the next time period, the method may further include: S301 to S304.

[0138] S301, obtaining data on training traffic usage behavior and training traffic usage;

[0139] S302, inputting the training traffic usage behavior data into an initial traffic determination model, and determining a target traffic usage amount corresponding to the training traffic usage behavior based on a relationship between the traffic usage behavior data and traffic usage amount;

[0140] S303, when the difference between the target traffic usage and the training traffic usage is not less than a set threshold, determining a traffic determination sub-model based on the training traffic usage behavior data, the difference between the target traffic usage and the training traffic usage, and the relationship between the model parameters;

[0141] S304: Obtain a flow determination model according to the initial flow determination model and the flow determination sub-model.

[0142] During model training, the initial model and sub-model are used to form the final traffic determination model. The initial model ensures global stability, and the sub-model optimizes local accuracy, thereby improving the model prediction accuracy.

[0143] In some embodiments, step S303, when the difference between the target traffic usage and the training traffic usage is not less than a set threshold, determining the traffic determination sub-model based on the relationship between the training traffic usage behavior data, the difference, and the model parameters, may include:

[0144] Calculate the difference between target traffic usage and training traffic usage;

[0145] Compute negative gradients and second-order derivatives from the differences;

[0146] Train the sub-model with the negative gradient as the response variable.

[0147] In some embodiments, the negative gradient and the second derivative are calculated as follows:

[0148]

[0149] in, is the negative gradient of the i-th sample at the t-th iteration, is the preset loss function, is the predicted value of the i-th sample at the t-th iteration, y i is the data of the training traffic usage behavior of the i-th sample, is the second-order derivative of the i-th sample at the t-th iteration.

[0150] Figure 4 An apparatus 400 for determining user traffic usage behavior provided by an embodiment of the present application is shown. The apparatus may include:

[0151] Acquisition module 401 is used to obtain the user's traffic usage behavior data for the current time period and the historical behavior change probability, where the historical behavior change probability is the probability that the user's traffic usage behavior will shift from one period in the historical time period to the next;

[0152] Determination module 402, for respectively calculating the distance between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors, and determining the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distance and the probability of the user's usage behavior;

[0153] Calculation module 403, used to calculate the user's target usage behavior probability in the next time period of the current time period based on the user's usage behavior probability and the historical behavior change probability;

[0154] The determination module 402 is further configured to determine the user's traffic usage behavior in the next time period based on the probability of the user's target usage behavior in the next time period of the current time period.

[0155] In some embodiments, the calculation module 403 is further configured to calculate the sum of the distances between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors to obtain an overall distance;

[0156] The calculation module 403 is further configured to calculate the ratio of the distance between the data of the traffic usage behavior and the data of each preset traffic usage behavior to the overall distance, thereby obtaining a normalized result of the distance between the data of the traffic usage behavior and the data of each preset traffic usage behavior;

[0157] The determination module 402 is further configured to determine the user usage behavior probability that the traffic usage behavior data belongs to the preset behavior corresponding to each preset traffic usage behavior data based on the relationship between the normalization result and the user usage behavior probability.

[0158] In some embodiments, the determination module 402 is further configured to determine a probability of a first traffic usage behavior of the user in the next time period based on the user usage behavior probability and the historical behavior change probability;

[0159] The determination module 402 determines the user target usage behavior probability of the user in the next time period according to the first traffic usage behavior probability, the traffic usage behavior data of the current time period and the preset weight vector.

[0160] In some embodiments, the apparatus 400 for determining user traffic usage behavior may further include:

[0161] The acquisition module 401 is further used to obtain data on training traffic usage behavior;

[0162] an allocation module, configured to select data of training traffic usage behaviors that meet preset conditions as initial cluster centers, and allocate the data of training traffic usage behaviors other than the data of training traffic usage behaviors corresponding to the initial cluster centers to the cluster where the nearest initial cluster center is located based on the distance between the data of training traffic usage behaviors other than the data of training traffic usage behaviors corresponding to the initial cluster centers and the data of the initial cluster centers, thereby obtaining a plurality of clusters;

[0163] The update module is used to calculate the feature mean of all training traffic usage behavior data in each cluster and update the cluster center according to the feature mean;

[0164] The determination module 402 is further configured to determine, when the cluster center no longer changes, that the data of the traffic usage behavior corresponding to the cluster center is the preset traffic usage behavior data.

[0165] In some embodiments, the determination module 402 is also used to input the data of the user's traffic usage behavior in the current time period into the traffic determination model, and use the traffic determination model to determine the traffic usage value in the next time period based on the data of the traffic usage behavior in the current time period, the pre-trained parameters of the traffic determination model and the relationship between the traffic usage value.

[0166] In some embodiments, the acquisition module 401 is further configured to acquire data on training traffic usage behavior and training traffic usage;

[0167] The determination module 402 is further configured to input the training traffic usage behavior data into the initial traffic determination model, and determine the target traffic usage corresponding to the training traffic usage behavior based on the relationship between the traffic usage behavior data and the traffic usage;

[0168] The determination module 402 is further configured to determine a traffic determination sub-model based on the relationship between the training traffic usage behavior data, the difference between the target traffic usage and the training traffic usage, and the model parameters when the difference between the target traffic usage and the training traffic usage is not less than a set threshold;

[0169] The determination module 402 is further configured to obtain a flow determination model according to the initial flow determination model and the flow determination sub-model.

[0170] Figure 4 The various modules in the device shown can be implemented Figure 1 Each step in the embodiment achieves the corresponding technical effect, which will not be described here for the sake of brevity.

[0171] Figure 5 A schematic diagram of the hardware structure of the terminal device provided in an embodiment of the present application is shown.

[0172] The terminal device may include a processor 501 and a memory 502 storing computer program instructions.

[0173] Specifically, the processor 501 may include a central processing unit (CPU) or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0174] The memory 502 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 502 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. In one example, the memory 502 may include a removable or non-removable (or fixed) medium, or the memory 502 may be a non-volatile solid-state memory. The memory 502 may be inside or outside the integrated gateway disaster recovery device.

[0175] In one example, the memory 502 may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical, or other physical / tangible memory storage device. Therefore, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method for determining user traffic usage behavior according to the present disclosure.

[0176] The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement Figure 1 The method for determining user traffic usage behavior in the illustrated embodiment.

[0177] In one example, the terminal device may further include a communication interface 503 and a bus 504. Figure 5 As shown, the processor 501 , the memory 502 , and the communication interface 503 are connected via a bus 504 and communicate with each other.

[0178] The communication interface 503 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0179] Bus 504 includes hardware, software or both, couples the parts of terminal equipment to each other.For example, but not limitation, bus may include Accelerated Graphics Port (AGP) or other graphics buses, Enhanced Industry Standard Architecture (EISA) bus, Front Side Bus (FSB), Hyper Transport (HT) interconnection, Industry Standard Architecture (ISA) bus, InfiniBand interconnection, Low Pin Count (LPC) bus, memory bus, Micro Channel Architecture (MCA) bus, Peripheral Component Interconnect (PCI) bus, PCI-Express (PCI-X) bus, Serial Advanced Technology Attachment (SATA) bus, Video Electronics Standards Association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 504 may include one or more buses. Although the present application embodiment describes and shows specific bus, the application considers any suitable bus or interconnection.

[0180] In addition, in conjunction with the method for determining user traffic usage behavior in the above embodiments, embodiments of the present application may provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any of the methods for determining user traffic usage behavior in the above embodiments is implemented.

[0181] An embodiment of the present application also provides a computer program product, including a computer program, which, when processed and executed, implements any one of the methods for determining user traffic usage behavior in the above embodiments.

[0182] It should be understood that the present application is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and illustrated as examples. However, the method process of the present application is not limited to the specific steps described and illustrated. Those skilled in the art can make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present application.

[0183] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of the present application are programs or text segments used to perform the required tasks. The programs or text segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memories (ROMs), flash memory, erasable read-only memories (EROMs), floppy disks, compact disc read-only memories (CD-ROMs), optical discs, hard disks, optical fiber media, radio frequency (RF) links, etc. The text segments can be downloaded via computer networks such as the Internet and intranets.

[0184] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps. In other words, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0185] Aspects of the present disclosure have been described above with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor or a field programmable logic circuit. It is also understood that each box in the block diagram and / or flowchart and the combination of the boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0186] The above description is only a specific embodiment of the present application. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the scope of protection of the present application.

Claims

1. A method for determining user traffic usage behavior, characterized in that: The method comprises: Obtaining data on the user's traffic usage behavior in the current time period and a historical behavior change probability, where the historical behavior change probability is the probability that the user's traffic usage behavior will shift from one period in the historical time period to the next; Calculating the distances between the data of the traffic usage behavior and data of multiple preset traffic usage behaviors respectively, and determining the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distances and the probability of user usage behavior; Calculate the user's target usage behavior probability for the next time period of the current time period based on the user's usage behavior probability and the historical behavior change probability; The user's traffic usage behavior in the next time period is determined based on the user's target usage behavior probability in the next time period of the current time period.

2. The method for determining user traffic usage behavior according to claim 1, characterized in that: The determining, based on the relationship between the distance and the probability of user usage behavior, the probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior includes: Calculate the sum of the distances between the data of the traffic usage behavior and the data of multiple preset traffic usage behaviors to obtain an overall distance; Calculating the ratio of the distance between the data of the traffic usage behavior and the data of each preset traffic usage behavior to the overall distance, respectively, to obtain a normalized result of the distance between the data of the traffic usage behavior and the data of each preset traffic usage behavior; According to the relationship between the normalization result and the probability of user usage behavior, the probability of user usage behavior that the data of traffic usage behavior belongs to the preset behavior corresponding to the data of each preset traffic usage behavior is determined.

3. The method for determining user traffic usage behavior according to claim 1, characterized in that: The calculating the user target usage behavior probability of the user in the next time period according to the user usage behavior probability and the historical behavior change probability includes: Determine the probability of the user's first traffic usage behavior in the next time period based on the user's usage behavior probability and the probability of historical behavior changes; The user target usage behavior probability of the user in the next time period is determined according to the first traffic usage behavior probability, the traffic usage behavior data of the current time period and the preset weight vector.

4. The method for determining user traffic usage behavior according to claim 1, characterized in that: Before respectively calculating the distances between the data of the traffic usage behavior and data of a plurality of preset traffic usage behaviors, and determining the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distances and the user usage behavior probabilities, the method further includes: Obtain data on training traffic usage behavior; The data of the training traffic usage behavior that meets the preset conditions are selected as the initial cluster center, and based on the distance between the data of the training traffic usage behavior other than the data of the training traffic usage behavior corresponding to the initial cluster center and the data of the initial cluster center, the data of the training traffic usage behavior other than the data of the training traffic usage behavior corresponding to the initial cluster center are assigned to the cluster where the nearest initial cluster center is located, thereby obtaining multiple clusters; Calculating the feature mean of all training traffic usage behavior data in each cluster, and updating the cluster center of the cluster according to the feature mean; When the cluster center no longer changes, the data of the traffic usage behavior corresponding to the cluster center is determined to be the data of the preset traffic usage behavior.

5. The method for determining user traffic usage behavior according to claim 1, characterized in that: After determining the traffic usage behavior corresponding to the user target usage behavior probability that is not less than a set threshold as the user's traffic usage behavior in the next time period, the method further includes: The data of the user's traffic usage behavior in the current time period is input into the traffic determination model, and the traffic determination model is used to determine the traffic usage value in the next time period based on the relationship between the data of the traffic usage behavior in the current time period, the pre-trained parameters of the traffic determination model and the traffic usage value.

6. The method for determining user traffic usage behavior according to claim 5, characterized in that: Before determining the traffic usage behavior corresponding to the user target usage behavior probability that is not less than a set threshold as the user's traffic usage behavior in the next time period, the method further includes: Obtain data on training traffic usage behavior and training traffic usage; Inputting the training traffic usage behavior data into an initial traffic determination model, and determining a target traffic usage amount corresponding to the training traffic usage behavior based on a relationship between the traffic usage behavior data and traffic usage amount; When the difference between the target traffic usage and the training traffic usage is not less than a set threshold, determining a traffic determination sub-model according to the relationship between the training traffic usage behavior data, the difference between the target traffic usage and the training traffic usage, and the model parameters; A flow determination model is obtained according to the initial flow determination model and the flow determination sub-model.

7. A device for determining user traffic usage behavior, characterized in that: The device comprises: An acquisition module is used to obtain the user's traffic usage behavior data for the current time period and the historical behavior change probability, wherein the historical behavior change probability is the probability that the user's traffic usage behavior will shift from one period in the historical time period to the next period; a determination module, configured to respectively calculate the distance between the data of the traffic usage behavior and the data of a plurality of preset traffic usage behaviors, and determine the usage behavior probability that the data of the traffic usage behavior belongs to each preset traffic usage behavior based on the relationship between the distance and the probability of user usage behavior; A calculation module, configured to calculate a user target usage behavior probability of the user in the next time period of the current time period based on the user usage behavior probability and the historical behavior change probability; The determination module is further used to determine the user's traffic usage behavior in the next time period based on the user's target usage behavior probability in the next time period of the current time period.

8. A terminal device, characterized in that: The device includes: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, it implements the method for determining user traffic usage behavior according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the method for determining user traffic usage behavior according to any one of claims 1 to 6.

10. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the method for determining user traffic usage behavior as described in any one of claims 1 to 6.

Images