Method and device for predicting complaining users

Abstract

An embodiment of the present invention provides a method and apparatus for predicting complaining users, wherein the method for predicting complaining users includes: obtaining historical user data within a preset time period; determining at least one complaint scenario from preset scenarios based on the historical user data, and establishing a complaint user classifier for each complaint scenario; wherein the user complaint rate of the complaint scenario is higher than the user complaint rate of the non-complaint scenario in the preset scenarios; obtaining user data to be processed; and processing the user data using the complaint user classifier corresponding to at least one complaint scenario to obtain potential complaining users. Since the complaint scenarios and the complaint user classifier for each complaint scenario are determined using historical user data within a preset time period, and thus potential complaining users are obtained, reliance on manual experience is avoided, thereby improving the accuracy and efficiency of complaining user prediction.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of data processing technology, and in particular to a method and device for predicting complaining users. Background Art

[0002] Large e-commerce companies can have hundreds of millions of users, but very few of them report complaints. Therefore, identifying users at risk of complaints from this massive user base is crucial.

[0003] Currently, a common method for identifying potential complaint users is based on business behavior. Based on business experience, the business behaviors of users with complaint risk are summarized. Furthermore, users with complaint risk are screened from a large user base based on their business behaviors. For example, a possible business behavior might be "order not shipped within three days of order placement, and customer service contacted more than twice." Users meeting this condition can be identified from a large user base as complaint risk users.

[0004] However, the aforementioned identification methods rely on manual business experience. The process of summarizing rules based on business experience is time-consuming and labor-intensive. Furthermore, the resulting business rules are limited and fail to effectively cover complaining users. Furthermore, the need to process data from all users results in low accuracy and efficiency in predicting complaining users. Summary of the Invention

[0005] The embodiments of the present invention provide a method and device for predicting complaining users, which improve the accuracy and efficiency of predicting complaining users.

[0006] In a first aspect, an embodiment of the present invention provides a method for predicting complaining users, comprising:

[0007] Get historical user data within a preset time period;

[0008] Determining at least one complaint scenario from preset scenarios based on the historical user data, and establishing a complaint user classifier for each complaint scenario; wherein the user complaint rate of the complaint scenario is higher than the user complaint rate of the non-complaint scenario in the preset scenarios;

[0009] Get the user data to be processed;

[0010] The user data is processed using a complaint user classifier corresponding to the at least one complaint scenario to obtain potential complaining users.

[0011] Optionally, determining at least one complaint scenario from preset scenarios based on the historical user data includes:

[0012] From the historical user data, a historical user data set for each of the preset scenarios is obtained;

[0013] Obtain the user complaint rate for each scenario based on the historical user data set for each scenario;

[0014] The user complaint rates corresponding to the preset scenarios are sorted in descending order, and the first N scenarios are determined as the at least one complaint scenario; wherein N is a positive integer.

[0015] Optionally, establishing a complaint user classifier for each complaint scenario includes:

[0016] According to the historical user data set of the complaint scenario, a user set of the complaint scenario is obtained; the user set includes all complaining users determined according to the historical user data set;

[0017] In the historical user data set of the complaint scenario, data of each user included in the user set is obtained to form the data set of the complaint scenario;

[0018] The data set of the complaint scenario is used as the input feature of the complaint user classifier, the user's complaint probability is used as the output feature of the complaint user classifier, the complaint user classifier is trained, and the complaint user classifier for the complaint scenario is established.

[0019] Optionally, the user set also includes some non-complaint users among all non-complaint users determined based on the historical user data set.

[0020] Optionally, the processing the user data using a complaint user classifier corresponding to the at least one complaint scenario to obtain potential complaining users includes:

[0021] Processing the user data using the complaint user classifiers corresponding to the at least one complaint scenario to obtain a complaint probability of each user in the user data in the at least one complaint scenario;

[0022] Obtaining a weight value of the at least one complaint scenario;

[0023] Obtaining a total complaint probability value for each user according to a weight value of the at least one complaint scenario and a complaint probability of each user in the at least one complaint scenario;

[0024] Potential complaining users are obtained according to the total complaint probability value of each user.

[0025] Optionally, processing the user data using the complaint user classifier corresponding to the at least one complaint scenario to obtain the complaint probability of each user in the user data in the at least one complaint scenario includes:

[0026] From the user data, obtain a user data set for each complaint scenario;

[0027] For each complaint scenario, the user data set of the complaint scenario is processed using the complaint user classifier of the complaint scenario to obtain the complaint probability of each user in the user data set in the complaint scenario.

[0028] Optionally, also include:

[0029] If the user data set of the first complaint scenario does not include the first user, the complaint probability of the first user in the first complaint scenario is determined to be 0.

[0030] Optionally, the user complaint rate of the complaint scenario is greater than or equal to a preset threshold.

[0031] In a second aspect, an embodiment of the present invention provides a device for predicting complaining users, comprising:

[0032] The first acquisition module is used to obtain historical user data within a preset time period;

[0033] a modeling module, configured to determine at least one complaint scenario from preset scenarios based on the historical user data, and establish a complaint user classifier for each of the complaint scenarios; wherein the user complaint rate of the complaint scenario is higher than the user complaint rate of the non-complaint scenario in the preset scenarios;

[0034] The second acquisition module is used to acquire user data to be processed;

[0035] The prediction module is used to process the user data using the complaint user classifier corresponding to the at least one complaint scenario to obtain potential complaining users.

[0036] Optionally, the modeling module is specifically used to:

[0037] From the historical user data, a historical user data set for each of the preset scenarios is obtained;

[0038] Obtain the user complaint rate for each scenario based on the historical user data set for each scenario;

[0039] The user complaint rates corresponding to the preset scenarios are sorted in descending order, and the first N scenarios are determined as the at least one complaint scenario; wherein N is a positive integer.

[0040] Optionally, the modeling module is specifically used to:

[0041] According to the historical user data set of the complaint scenario, a user set of the complaint scenario is obtained; the user set includes all complaining users determined according to the historical user data set;

[0042] In the historical user data set of the complaint scenario, data of each user included in the user set is obtained to form the data set of the complaint scenario;

[0043] The data set of the complaint scenario is used as the input feature of the complaint user classifier, the user's complaint probability is used as the output feature of the complaint user classifier, the complaint user classifier is trained, and the complaint user classifier for the complaint scenario is established.

[0044] Optionally, the user set also includes some non-complaint users among all non-complaint users determined based on the historical user data set.

[0045] Optionally, the prediction module is specifically used to:

[0046] Processing the user data using the complaint user classifiers corresponding to the at least one complaint scenario to obtain a complaint probability of each user in the user data in the at least one complaint scenario;

[0047] Obtaining a weight value of the at least one complaint scenario;

[0048] Obtaining a total complaint probability value for each user according to a weight value of the at least one complaint scenario and a complaint probability of each user in the at least one complaint scenario;

[0049] Potential complaining users are obtained according to the total complaint probability value of each user.

[0050] Optionally, the prediction module is specifically used to:

[0051] From the user data, obtain a user data set for each complaint scenario;

[0052] For each complaint scenario, the user data set of the complaint scenario is processed using the complaint user classifier of the complaint scenario to obtain the complaint probability of each user in the user data set in the complaint scenario.

[0053] Optionally, the prediction module is further configured to:

[0054] If the user data set of the first complaint scenario does not include the first user, the complaint probability of the first user in the first complaint scenario is determined to be 0.

[0055] Optionally, the user complaint rate of the complaint scenario is greater than or equal to a preset threshold.

[0056] In a third aspect, an embodiment of the present invention provides a device for predicting complaining users, comprising: a processor and a memory;

[0057] The memory is used to store instructions;

[0058] The processor is used to execute instructions stored in the memory to perform the method provided by any implementation of the first aspect of the present invention.

[0059] In a fourth aspect, an embodiment of the present invention provides a storage medium, comprising: a readable storage medium and a computer program, wherein the computer program is used to implement the method provided in any implementation manner of the first aspect of the present invention.

[0060] An embodiment of the present invention provides a method and device for predicting complaint users, which determines complaint scenarios and complaint user classifiers for each complaint scenario through historical user data within a preset time period, thereby improving the accuracy of determining complaint scenarios and establishing complaint user classifiers. Therefore, by processing the user data to be processed according to the complaint user classifier corresponding to each complaint scenario, potential complaint users can be obtained, avoiding reliance on manual experience, and improving the accuracy and efficiency of complaint user prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0062] Figure 1 A flow chart of a method for predicting complaining users provided by an embodiment of the present invention;

[0063] Figure 2 Another flow chart of the method for predicting complaining users provided by an embodiment of the present invention;

[0064] Figure 3 A schematic diagram illustrating the principle of obtaining the total probability of user complaints provided by an embodiment of the present invention;

[0065] Figure 4 A schematic diagram of the structure of a device for predicting complaining users provided by an embodiment of the present invention;

[0066] Figure 5 Another structural diagram of the device for predicting complaining users provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0067] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0068] Figure 1 A flow chart of the method for predicting complaining users provided by an embodiment of the present invention. The method for predicting complaining users provided by this embodiment can be executed by a prediction device for complaining users. Figure 1 As shown, the method for predicting complaining users provided in this embodiment may include:

[0069] S101: Obtain historical user data within a preset time period.

[0070] Among them, historical user data is data related to the user's business behavior on the e-commerce platform. This embodiment does not limit the specific content of historical user data. Optionally, historical user data may include but is not limited to at least one of the following: user's product browsing data, order data, logistics data, comment data, etc. on the e-commerce platform. Product browsing data can reflect information such as the type, price, browsing time, and number of views of the products browsed by the user. Order data can reflect information such as the type, price, quantity, etc. of the products of which the user's transaction is successful or failed. Comment data may include but is not limited to user feedback on the product, communication information between the user and customer service, complaint information, etc. Logistics data can reflect information such as the shipping time, transportation time, and delivery time of the product during the user's purchase process.

[0071] In this embodiment, the specific value of the preset time length is not limited.

[0072] S102: Determine at least one complaint scenario from preset scenarios based on historical user data, and establish a complaint user classifier for each complaint scenario.

[0073] Among them, the user complaint rate in the complaint scenario is higher than the user complaint rate in the non-complaint scenario in the preset scenario.

[0074] Specifically, multiple preset scenarios are pre-set, and this embodiment does not limit the number and classification of preset scenarios. For example, preset scenarios may include but are not limited to: shipping scenarios, return scenarios, distribution scenarios, order scenarios, merchant service scenarios, and so on. Based on historical user data within a preset time period, at least one complaint scenario can be determined in multiple preset scenarios, and a complaint user classifier for each complaint scenario can be established. The complaint user classifier for a complaint scenario is used to output the probability of a user complaining in the complaint scenario. By determining complaint scenarios with higher user complaint rates and complaint user classifiers for each complaint scenario through historical user data, the accuracy of determining complaint scenarios and establishing complaint user classifiers is improved.

[0075] The user complaint rate for complaint scenarios is higher than that for non-complaint scenarios. This embodiment does not limit the number of complaint scenarios.

[0076] Optionally, to further improve the accuracy of determining the complaint scenario, the user complaint rate of the complaint scenario may be greater than or equal to a preset threshold. This embodiment does not limit the specific value of the preset threshold.

[0077] Optionally, the user complaint rate for each preset scenario may be the ratio of the number of complaints in the preset scenario to the sum of the number of complaints in all preset scenarios in historical user data, which is specifically expressed by Formula 1. k represents the number of complaints for the kth preset scenario, I k represents the user complaint rate for the kth preset scenario. T represents the number of preset scenarios, and T is an integer greater than 1.

[0078]

[0079] Optionally, the user complaint rate of each preset scenario may be the ratio of the number of complaining users in the preset scenario to the sum of the number of complaining users in all preset scenarios, which is specifically expressed by Formula 2. k represents the number of users who complained about the kth preset scenario, J k represents the user complaint rate for the kth preset scenario. T represents the number of preset scenarios, and T is an integer greater than 1.

[0080]

[0081] S103: Obtain user data to be processed.

[0082] The content of the user data to be processed can be found in the above-mentioned historical user data. The principle is similar and will not be repeated here.

[0083] S104: Process user data using a complaint user classifier corresponding to at least one complaint scenario to obtain potential complaining users.

[0084] It can be seen that the complaint user prediction method provided in this embodiment uses historical user data within a preset time period to determine the complaint scenario and the complaint user classifier for each complaint scenario, thereby improving the accuracy of determining the complaint scenario and establishing the complaint user classifier. Therefore, the user data to be processed is processed according to the complaint user classifier corresponding to each complaint scenario, and potential complaint users can be obtained, avoiding reliance on manual experience, thereby improving the accuracy and efficiency of complaint user prediction.

[0085] Optionally, in S102, determining at least one complaint scenario from preset scenarios based on historical user data may include:

[0086] In the historical user data, a historical user data set for each scenario in the preset scenario is obtained.

[0087] Based on the historical user data set of each scenario, obtain the user complaint rate of each scenario.

[0088] The user complaint rates corresponding to the preset scenarios are sorted in descending order, and the top N scenarios are determined as at least one complaint scenario, where N is a positive integer.

[0089] Specifically, each preset scenario corresponds to a historical user data set, which includes data related to the preset scenario from historical user data within a preset time period. Based on the historical user data set for each preset scenario, the user complaint rate for each preset scenario is obtained. By sorting the user complaint rates of multiple preset scenarios, the complaint scenario can be determined.

[0090] The following example illustrates this. Assume that the pre-set scenarios include shipping, returns, logistics, orders, and merchant service. N is 2. Based on historical user data from a pre-set time period, obtain a collection of historical user data for each scenario. The user complaint rates for shipping, returns, logistics, orders, and merchant service are 10%, 40%, 25%, 5%, and 20%, respectively. After sorting the user complaint rates from highest to lowest, it can be determined that the two most common complaint scenarios are returns and logistics.

[0091] By determining the top N scenarios ranked by user complaint rate as complaint scenarios, the accuracy of determining complaint scenarios is further improved.

[0092] Optionally, in S102, establishing a complaint user classifier for each complaint scenario may include:

[0093] According to the historical user data set of the complaint scenario, a user set of the complaint scenario is obtained. The user set includes all complaining users determined according to the historical user data set.

[0094] In the historical user data set of the complaint scenario, data of each user included in the user set is obtained to form a data set of the complaint scenario.

[0095] The data set of complaint scenarios is used as the input feature of the complaint user classifier, and the user's complaint probability is used as the output feature of the complaint user classifier. The complaint user classifier is trained to establish a complaint user classifier for complaint scenarios.

[0096] The following uses the logistics scenario as an example to illustrate the complaint scenario.

[0097] Assume that, based on the historical user data set for the logistics scenario, the user set obtained for the logistics scenario includes 200 complaining users. Furthermore, based on the historical user data set for the logistics scenario, it can be determined that there are 1000 non-complaining users in the logistics scenario. Within the historical user data set for the logistics scenario, data on 200 complaining users is obtained to form the data set for the logistics scenario. The data of these 200 complaining users is used as the input features of a complaint user classifier for the logistics scenario, and the complaint probability of each complaining user is used as the output feature of the complaint user classifier. The complaint user classifier for the logistics scenario is trained to establish a complaint user classifier for the complaint scenario.

[0098] The complaint user classifier for each complaint scenario is trained using the data of the complaining users in that scenario. Training is performed separately for each complaint scenario, thereby improving the accuracy of the complaint user classifier for each complaint scenario.

[0099] Optionally, the user set of the complaint scenario may also include some non-complaint users among all non-complaint users determined based on the historical user data set.

[0100] Taking the aforementioned logistics scenario as an example, let's determine that there are 1,000 non-complaining users based on the historical user data set for the logistics scenario. Then, in addition to the 200 complaining users, the user set for the logistics scenario may also include some of the 1,000 non-complaining users. This embodiment does not limit the specific number of non-complaining users included.

[0101] Optionally, if there are N complaint scenarios, where N is a positive integer, the number of users included in the user set of the kth complaint scenario can be p k +r×(pool k -p k ). Among them, p k represents the number of users who complained in the kth complaint scenario, pool krepresents the number of all users in the kth complaint scenario, including both complaining and non-complaining users. r represents the extraction probability, which is greater than or equal to 0 and less than 1. It should be noted that for N complaint scenarios, the value of r can be the same or different for different complaint scenarios.

[0102] This embodiment provides a method for predicting complaining users, comprising: obtaining historical user data within a preset time period, determining at least one complaint scenario from preset scenarios based on the historical user data, establishing a complaint user classifier for each complaint scenario, obtaining user data to be processed, and processing the user data using the complaint user classifier corresponding to at least one complaint scenario to obtain potential complaining users. The method for predicting complaining users provided in this embodiment uses historical user data within a preset time period to determine complaint scenarios and complaint user classifiers for each complaint scenario, and then obtains potential complaining users based on the complaint user classifiers for each complaint scenario, avoiding reliance on manual experience and improving the accuracy and efficiency of complaint user prediction.

[0103] Figure 2 Another flow chart of the method for predicting complaint users provided by an embodiment of the present invention. Figure 1 Based on the embodiment shown, a specific implementation of S104 is provided. Figure 2 As shown, processing user data using a complaint user classifier corresponding to at least one complaint scenario to obtain potential complaining users may include:

[0104] S201. Process user data using a complaint user classifier corresponding to at least one complaint scenario to obtain a complaint probability of each user in the user data in at least one complaint scenario.

[0105] S202. Obtain a weight value of at least one complaint scenario.

[0106] S203. Obtain a total complaint probability value for each user based on a weight value of at least one complaint scenario and the complaint probability of each user in at least one complaint scenario.

[0107] S204: Obtain potential complaining users based on the total complaint probability of each user.

[0108] The following combination Figure 3 An exemplary description is given. Figure 3A schematic diagram of the principle of obtaining the total value of a user's complaint probability provided by an embodiment of the present invention. Assume that the number of complaint scenarios is N, and N is a positive integer. The complaint user classifiers for the N complaint scenarios are respectively referred to as complaint user classifiers 1 to complaint user classifiers N. After obtaining the user data to be processed, the user data is processed using complaint user classifiers 1 to complaint user classifiers N. For each user in the user data, complaint user classifiers 1 to complaint user classifiers N respectively output the complaint probability of the user in the corresponding complaint scenario. For example, in Figure 3 For a user, using complaint user classifier 1, we can obtain the user's complaint probability P1 in the first complaint scenario, using complaint user classifier 2, we can obtain the user's complaint probability P2 in the second complaint scenario, and so on. Using complaint user classifier N, we can obtain the user's complaint probability P in the Nth complaint scenario. N Then, according to the weight values ​​of N complaint scenarios and the complaint probability of each user in N complaint scenarios, the total complaint probability of each user is obtained. Figure 3 In the example, the weight values ​​of N complaint scenarios are represented as w1~w N For a user, the total probability of the user's complaint can be expressed as p1*w1+p2*w2+…+p N *w N After obtaining the total complaint probability value of each user in the user data, potential complaining users can be obtained based on the total complaint probability value of each user.

[0109] It can be seen that since the complaint user classifier corresponding to at least one complaint scenario is obtained based on historical data, it has a high accuracy rate. By processing the user data to be processed through the complaint user classifier corresponding to at least one complaint scenario, the total complaint probability of each user in the user data to be processed can be obtained, and then potential complaint users can be obtained, avoiding reliance on manual experience and improving the accuracy and efficiency of obtaining potential complaint users.

[0110] Optionally, in S201, processing the user data using a complaint user classifier corresponding to at least one complaint scenario to obtain the complaint probability of each user in the user data in at least one complaint scenario may include:

[0111] In the user data, obtain the user data set for each complaint scenario.

[0112] For each complaint scenario, the complaint user classifier of the complaint scenario is used to process the user data set of the complaint scenario to obtain the complaint probability of each user in the user data set in the complaint scenario.

[0113] Specifically, each complaint scenario corresponds to a user data set, which includes data related to the complaint scenario in the user data to be processed. By processing the user data set for each complaint scenario using the complaint user classifier for that complaint scenario, the complaint probability of each user in each complaint scenario can be obtained.

[0114] Optionally, the method for predicting complaining users provided in this embodiment may further include:

[0115] If the user data set of the first complaint scenario does not include the first user, the complaint probability of the first user in the first complaint scenario is determined to be 0.

[0116] For example, for user A, if the user data set of the shipping scenario does not include user A, the probability of user A making a complaint in the shipping scenario is 0.

[0117] This embodiment provides a method for predicting complaining users, specifically providing an implementation method for identifying potential complaining users using a complaint user classifier corresponding to at least one complaint scenario. By processing the user data to be processed using the complaint user classifier corresponding to at least one complaint scenario, the total complaint probability for each user in the processed user data can be obtained, thereby identifying potential complaining users. This avoids reliance on manual experience, improves the accuracy of identifying potential complaining users, and improves the efficiency of processing potential complaining users.

[0118] Figure 4 A schematic diagram of the structure of the prediction device for complaint users provided in an embodiment of the present invention. The prediction device for complaint users provided in this embodiment is used to execute the prediction method for complaint users provided in an embodiment of the present invention. Figure 4 As shown, the prediction device for complaining users provided in this embodiment may include:

[0119] The first acquisition module 41 is used to acquire historical user data within a preset time period;

[0120] a modeling module 42 for determining at least one complaint scenario from preset scenarios based on the historical user data and establishing a complaining user classifier for each of the complaint scenarios; wherein the user complaint rate of the complaint scenario is higher than the user complaint rate of the non-complaint scenario in the preset scenarios;

[0121] A second acquisition module 43 is used to acquire user data to be processed;

[0122] The prediction module 44 is configured to process the user data using a complaint user classifier corresponding to the at least one complaint scenario to obtain potential complaining users.

[0123] Optionally, the modeling module 42 is specifically configured to:

[0124] From the historical user data, a historical user data set for each of the preset scenarios is obtained;

[0125] Obtain the user complaint rate for each scenario based on the historical user data set for each scenario;

[0126] The user complaint rates corresponding to the preset scenarios are sorted in descending order, and the first N scenarios are determined as the at least one complaint scenario; wherein N is a positive integer.

[0127] Optionally, the modeling module 42 is specifically configured to:

[0128] According to the historical user data set of the complaint scenario, a user set of the complaint scenario is obtained; the user set includes all complaining users determined according to the historical user data set;

[0129] In the historical user data set of the complaint scenario, data of each user included in the user set is obtained to form the data set of the complaint scenario;

[0130] The data set of the complaint scenario is used as the input feature of the complaint user classifier, the user's complaint probability is used as the output feature of the complaint user classifier, the complaint user classifier is trained, and the complaint user classifier for the complaint scenario is established.

[0131] Optionally, the user set also includes some non-complaint users among all non-complaint users determined based on the historical user data set.

[0132] Optionally, the prediction module 44 is specifically configured to:

[0133] Processing the user data using the complaint user classifiers corresponding to the at least one complaint scenario to obtain a complaint probability of each user in the user data in the at least one complaint scenario;

[0134] Obtaining a weight value of the at least one complaint scenario;

[0135] Obtaining a total complaint probability value for each user according to a weight value of the at least one complaint scenario and a complaint probability of each user in the at least one complaint scenario;

[0136] Potential complaining users are obtained according to the total complaint probability value of each user.

[0137] Optionally, the prediction module 44 is specifically configured to:

[0138] From the user data, obtain a user data set for each complaint scenario;

[0139] For each complaint scenario, the user data set of the complaint scenario is processed using the complaint user classifier of the complaint scenario to obtain the complaint probability of each user in the user data set in the complaint scenario.

[0140] Optionally, the prediction module 44 is further configured to:

[0141] If the user data set of the first complaint scenario does not include the first user, the complaint probability of the first user in the first complaint scenario is determined to be 0.

[0142] Optionally, the user complaint rate of the complaint scenario is greater than or equal to a preset threshold.

[0143] The prediction device for complaining users provided in this embodiment is used to execute the prediction method for complaining users provided in the embodiment of the present invention. The technical principles are similar and will not be repeated here.

[0144] Figure 5 Another structural diagram of the device for predicting complaint users provided by an embodiment of the present invention. The device for predicting complaint users provided by this embodiment is used to execute the method for predicting complaint users provided by an embodiment of the present invention. Figure 5 As shown, the information processing device provided in this embodiment may include: a processor 51 and a memory 52. ​​The memory 52 is used to store program instructions. The processor 51 is used to call the program instructions stored in the memory 52 to implement the method for predicting complaining users provided in this embodiment of the present invention. The technical principles are similar and will not be repeated here.

[0145] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0146] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, rather than to limit them. Although the embodiments of the present invention have been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting complaining users, characterized in that: include: Get historical user data within a preset time period; Determine at least one complaint scenario from preset scenarios based on the historical user data, and establish a complaint user classifier for each complaint scenario; wherein the user complaint rate of the complaint scenario is higher than the user complaint rate of the non-complaint scenario in the preset scenarios, and the complaint user classifier is a classifier established by training using the data set of the complaint scenario as input features and the user's complaint probability as output features; Get the user data to be processed; Processing the user data using the complaint user classifiers corresponding to the at least one complaint scenario to obtain a complaint probability of each user in the user data in the at least one complaint scenario; Obtaining a weight value of the at least one complaint scenario; Obtaining a total complaint probability value for each user according to a weight value of the at least one complaint scenario and a complaint probability of each user in the at least one complaint scenario; Potential complaining users are obtained according to the total complaint probability value of each user.

2. The method according to claim 1, characterized in that Determining at least one complaint scenario from preset scenarios based on the historical user data includes: From the historical user data, a historical user data set for each of the preset scenarios is obtained; Obtain the user complaint rate for each scenario based on the historical user data set for each scenario; The user complaint rates corresponding to the preset scenarios are sorted in descending order, and the first N scenarios are determined as the at least one complaint scenario; wherein N is a positive integer.

3. The method according to claim 2, characterized in that The step of establishing a complaint user classifier for each complaint scenario includes: According to the historical user data set of the complaint scenario, a user set of the complaint scenario is obtained; the user set includes all complaining users determined according to the historical user data set; In the historical user data set of the complaint scenario, data of each user included in the user set is obtained to form the data set of the complaint scenario; The data set of the complaint scenario is used as the input feature of the complaint user classifier, the user's complaint probability is used as the output feature of the complaint user classifier, the complaint user classifier is trained, and the complaint user classifier for the complaint scenario is established.

4. The method according to claim 3, characterized in that The user set also includes some non-complaint users among all non-complaint users determined based on the historical user data set.

5. The method according to claim 4, characterized in that The processing of the user data by using the complaint user classifier corresponding to the at least one complaint scenario to obtain the complaint probability of each user in the user data in the at least one complaint scenario includes: From the user data, obtain a user data set for each complaint scenario; For each complaint scenario, the user data set of the complaint scenario is processed using the complaint user classifier of the complaint scenario to obtain the complaint probability of each user in the user data set in the complaint scenario.

6. A device for predicting complaining users, characterized in that: include: The first acquisition module is used to obtain historical user data within a preset time period; a modeling module for determining at least one complaint scenario from preset scenarios based on the historical user data and establishing a complaint user classifier for each complaint scenario; wherein the user complaint rate of the complaint scenario is higher than the user complaint rate of the non-complaint scenario in the preset scenarios, and the complaint user classifier is a classifier established by training using the data set of the complaint scenario as input features and the user complaint probability as output features; The second acquisition module is used to acquire user data to be processed; a prediction module, configured to process the user data using a complaint user classifier corresponding to the at least one complaint scenario to obtain potential complaining users; The prediction module is specifically configured to: process the user data using the complaint user classifier corresponding to the at least one complaint scenario, and obtain the complaint probability of each user in the user data in the at least one complaint scenario; Obtaining a weight value of the at least one complaint scenario; Obtaining a total complaint probability value for each user according to a weight value of the at least one complaint scenario and a complaint probability of each user in the at least one complaint scenario; Potential complaining users are obtained according to the total complaint probability value of each user.

7. The device according to claim 6, characterized in that The modeling module is specifically used for: From the historical user data, a historical user data set for each of the preset scenarios is obtained; Obtain the user complaint rate for each scenario based on the historical user data set for each scenario; The user complaint rates corresponding to the preset scenarios are sorted in descending order, and the first N scenarios are determined as the at least one complaint scenario; wherein N is a positive integer.

8. The device according to claim 7, characterized in that The modeling module is specifically used for: According to the historical user data set of the complaint scenario, a user set of the complaint scenario is obtained; the user set includes all complaining users determined according to the historical user data set; In the historical user data set of the complaint scenario, data of each user included in the user set is obtained to form the data set of the complaint scenario; The data set of the complaint scenario is used as the input feature of the complaint user classifier, the user's complaint probability is used as the output feature of the complaint user classifier, the complaint user classifier is trained, and the complaint user classifier for the complaint scenario is established.

9. A device for predicting complaining users, characterized in that: include: processor and memory; The memory is used to store instructions; The processor is configured to execute instructions stored in the memory to perform the method according to any one of claims 1 to 5.

10. A storage medium, characterized in that: include: A readable storage medium and a computer program, wherein the computer program is used to implement the method according to any one of claims 1 to 5.

Images