Low-net-age complaint user identification method and device and readable storage medium

Abstract

The invention provides a low-network-age complaint user identification method and device and a readable storage medium. The method comprises the steps of obtaining network access data and service data of a user; inputting the network access data and the service data of the user into a trained neural network model, identifying whether the user is a low-network-age complaint user through the trained neural network model, and obtaining an identification result of the user, the trained neural network model is obtained by training based on network access data and service data of a plurality of historical users. According to the method, the device and the readable storage medium, the problem that no related scheme for identifying the low-net-age complaint user exists in the prior art can be solved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device and readable storage medium for identifying users with low Internet age complaints. Background Art

[0002] With the development of operators' business, a large number of users join the network every day, and the quality of these users is uneven. There are even fake users who take advantage of the free quota subsidized by the operator, complain or cancel the account after use, and do not make substantial recharges, or complain and cancel the account soon after joining the network, in order to obtain the operator's commission. This will cause great losses to the operator. Fake users occupy valuable number card resources without bringing substantial income. After the complaint, the service cost of the operator will increase, and the number of complaints will increase. The most important thing is that the occupation of such public service resources will also affect the service experience of normal users.

[0003] In the process of research and practice of existing technologies, the inventors found that most of the fake users are users with low Internet age who complain. Such users usually do not bring substantial income to the company, but require the company to pay additional costs, which brings obvious troubles to the customer service department. If these users can be identified in advance using the model, proactive intervention can be carried out to reduce the company's losses. However, there is currently no relevant solution to specifically identify users with low Internet age who complain. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method, device and readable storage medium for identifying users with low Internet age complaints in view of the above-mentioned deficiencies in the prior art, so as to solve the problem that the prior art has no relevant solutions for identifying users with low Internet age complaints.

[0005] In a first aspect, the present invention provides a method for identifying users with low Internet age complaints, comprising:

[0006] Obtain the user's network access data and business data;

[0007] The network access data and service data of the user are input into a trained neural network model, and the trained neural network model is used to identify whether the user is a low-network age complaint user to obtain an identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users.

[0008] Furthermore, before inputting the network access data and service data of the user into the trained neural network model, identifying whether the user is a low-network age complaint user through the trained neural network model, and obtaining the identification result of the user, the method also includes:

[0009] Obtaining a training data set according to the network access data and service data of the plurality of historical users;

[0010] Marking whether each historical user in the training data set is a low-age Internet complaint user to obtain a marked training data set;

[0011] The pre-selected neural network model is trained using the labeled training data set to obtain the trained neural network model.

[0012] Furthermore, obtaining a training data set according to the network access data and service data of the plurality of historical users specifically includes:

[0013] According to the business data and business knowledge of the multiple historical users, key features related to the complaint behavior in the business data of the multiple historical users are selected, and new features of the multiple historical users are constructed;

[0014] The network access data, service data, key features related to the complaint behavior and new features of the multiple historical users are used as the training data set.

[0015] Furthermore, the neural network model is one of the following: a multi-layer perceptron MLP, a convolutional neural network CNN, or a recurrent neural network RNN.

[0016] Furthermore, the network access data includes: network access time, and the service data includes: development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills;

[0017] The inputting the network access data and service data of the user into the trained neural network model, identifying whether the user is a low-network age complaint user through the trained neural network model, and obtaining the identification result of the user specifically includes:

[0018] The user's network access time, development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills are input into the trained neural network model, and the neural network model obtains the user's key features related to the complaint behavior and the user's new features based on the user's service usage, historical complaint records, and real-time bills. The user's first complaint time is predicted based on the user's development channel, account opening category, product identification, whether to re-access the network, the key features related to the complaint behavior, and the new features. Based on the user's first complaint time and the user's network access time, it is identified whether the user is a low-network-age complaint user.

[0019] Furthermore, the step of identifying whether the user is a low-network age complaint user based on the time of the user's first complaint and the time of the user's network access specifically includes:

[0020] If the difference between the time of the user's first complaint and the time of the user's network access is less than a preset time period, the user is identified as a low-network-age complaint user.

[0021] Furthermore, the network access data and service data of the user are input into the trained neural network model, and the trained neural network model is used to identify whether the user is a user with low network age complaint. After obtaining the identification result of the user, the method further includes:

[0022] If the identification result of the user indicates that the user is a low-age Internet complaint user, manual intervention is performed on the user.

[0023] In a second aspect, the present invention provides a device for identifying users with low Internet age complaints, comprising:

[0024] The acquisition module is used to obtain the user's network access data and service data;

[0025] An identification module is connected to the acquisition module and is used to input the user's network access data and service data into a trained neural network model, identify whether the user is a low-network age complaint user through the trained neural network model, and obtain an identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users.

[0026] In a third aspect, the present invention provides a device for identifying users with low Internet age complaint, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the method for identifying users with low Internet age complaint described in the first aspect.

[0027] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for identifying users with low Internet age complaints as described in the first aspect above is implemented.

[0028] The method, device and readable storage medium for identifying users with low network age complaints provided by the present invention first obtain the user's network access data and service data; then input the user's network access data and service data into a trained neural network model, and use the trained neural network model to identify whether the user is a user with low network age complaints, and obtain the identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users. The present invention uses a trained neural network model to identify users with low network age complaints based on the user's network access data and service data, so that operators can intervene in advance to prevent users with low network age complaints from causing losses to operators, and curb such situations from happening at the source, solving the problem that the prior art has no relevant solutions for how to identify users with low network age complaints. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a flow chart of a method for identifying users with low Internet age complaints according to Embodiment 1 of the present invention;

[0030] Figure 2 This is a schematic diagram of the structure of a device for identifying users with low Internet age complaints according to Embodiment 2 of the present invention;

[0031] Figure 3 This is a structural diagram of a device for identifying users with low Internet age complaints according to Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the technical solution of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0033] It should be understood that the specific embodiments and drawings described herein are only used to explain the present invention rather than to limit the present invention.

[0034] It can be understood that, in the absence of conflict, the various embodiments of the present invention and the various features in the embodiments can be combined with each other.

[0035] It can be understood that, for the convenience of description, the drawings of the present invention only show the parts related to the present invention, while the parts irrelevant to the present invention are not shown in the drawings.

[0036] It can be understood that each unit and module involved in the embodiments of the present invention may correspond to only one physical structure, or may be composed of multiple physical structures, or multiple units and modules may be integrated into one physical structure.

[0037] It can be understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of the present invention may occur in an order different from that marked in the drawings.

[0038] It is understood that the flowcharts and block diagrams of the present invention illustrate the possible architectures, functions, and operations of the systems, devices, equipment, and methods according to the various embodiments of the present invention. Each box in the flowchart or block diagram may represent a unit, module, program segment, or code, which contains executable instructions for implementing the specified functions. Moreover, each box or combination of boxes in the block diagram and flowchart may be implemented by a hardware-based system that implements the specified functions, or may be implemented by a combination of hardware and computer instructions.

[0039] It can be understood that the units and modules involved in the embodiments of the present invention can be implemented by software or hardware. For example, the units and modules can be located in a processor.

[0040] Embodiment 1:

[0041] This embodiment provides a method for identifying low-age online complaint users. Figure 1 As shown, the method includes:

[0042] Step S101: Acquire the user's network access data and service data.

[0043] It should be noted that the network access data includes: network access time, etc., and the business data includes: development channels, account opening categories, product identification, whether to re-access the network, service usage, historical complaint records, real-time bills, etc.

[0044] Step S102: input the network access data and service data of the user into a trained neural network model, identify whether the user is a low-network age complaint user through the trained neural network model, and obtain an identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users.

[0045] In this embodiment, a neural network model is trained based on the network access data and service data of multiple historical users, so that the trained neural network model can identify in advance whether a user is a low-network age complaint user.

[0046] In an optional embodiment, before inputting the network access data and service data of the user into a trained neural network model, identifying whether the user is a low-network age complaint user through the trained neural network model, and obtaining the identification result of the user, the method further includes:

[0047] Obtaining a training data set according to the network access data and service data of the plurality of historical users;

[0048] Marking whether each historical user in the training data set is a low-age Internet complaint user to obtain a marked training data set;

[0049] The pre-selected neural network model is trained using the labeled training data set to obtain the trained neural network model.

[0050] Specifically, the network access data and service data of multiple historical users are obtained, and a training data set is obtained based on the network access data and service data of each historical user among the multiple historical users. The training data set is annotated to determine which historical users are low-network age complaint users. For example, for each historical user, a label is created for the historical user based on whether the historical user is a low-network age complaint user, such as "complaint" or "no complaint", a neural network model is selected, and the number of network layers, the number of neurons in each layer, the activation function and other parameters are determined, wherein the neural network model is one of the following: multilayer perceptron (MLP), convolutional neural network (CNN), recurrent neural network (RNN). The pre-selected neural network model is trained using the annotated training data set to obtain a trained neural network model.

[0051] In an optional embodiment, obtaining a training data set according to the network access data and service data of the plurality of historical users specifically includes:

[0052] According to the business data and business knowledge of the multiple historical users, key features related to the complaint behavior in the business data of the multiple historical users are selected, and new features of the multiple historical users are constructed;

[0053] The network access data, service data, key features related to the complaint behavior and new features of the multiple historical users are used as the training data set.

[0054] It should be noted that part of the information in the user's business data can be directly used as features of the training data set, such as development channels, account opening categories, product identification, and whether to re-join the network, while another part of the information in the user's business data needs to further select key features related to the complaint behavior and construct new features based on the other part of the information as features of the training data set.

[0055] It should be noted that the historical user access data includes not only the access time, but also the complaint time.

[0056] Specifically, based on business knowledge, the business data of multiple historical users are screened, and key features related to complaint behavior in the business data of multiple historical users are selected, such as network access time, service usage frequency, etc. And new features of multiple historical users are constructed based on the business data of multiple historical users, such as user activity, service failure rate, etc., where user activity is the number of historical complaints. When the number of historical complaints is less than the first preset number, it means inactivity, when it is greater than the first preset number and less than the second preset number, it means general, and when it is greater than the second preset number, it means activity. In order to better capture the user's complaint tendency. The network access time and complaint time in the network access data of multiple historical users, the development channel, account opening category, product identification and whether to re-access the network in the business data, the network access time and service usage frequency in the key features related to complaint behavior, and the user activity and service failure rate in the new features are used as training data sets.

[0057] In an optional embodiment, the inputting the network access data and service data of the user into a trained neural network model, identifying whether the user is a low-network age complaint user through the trained neural network model, and obtaining the identification result of the user specifically includes:

[0058] The user's network access time, development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills are input into the trained neural network model, and the neural network model obtains the user's key features related to the complaint behavior and the user's new features based on the user's service usage, historical complaint records, and real-time bills. The user's first complaint time is predicted based on the user's development channel, account opening category, product identification, whether to re-access the network, the key features related to the complaint behavior, and the new features. Based on the user's first complaint time and the user's network access time, it is identified whether the user is a low-network-age complaint user.

[0059] Specifically, the user's network access time, development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills are input into the trained neural network model, and the trained neural network model obtains the user's network access time and service usage frequency, as well as the user's user activity and service failure rate based on the user's service usage, historical complaint records, and real-time bills, and predicts the user's first complaint time based on the user's development channel, account opening category, product identification, whether to re-access the network, network access time, service usage frequency, user activity, and service failure rate. The trained neural network model identifies whether the user is a low-network-age complaint user based on the predicted first complaint time and the user's network access time, and outputs the user's identification result.

[0060] Specifically, based on the user's first complaint time and the user's network access time, the specific method for identifying whether the user is a low-network-age complaint user can be: if the difference between the user's first complaint time and the user's network access time is less than a preset time period (such as 6 months), then the user is identified as a low-network-age complaint user.

[0061] In an optional embodiment, the network access data and service data of the user are input into a trained neural network model, and the trained neural network model is used to identify whether the user is a low-network age complaint user. After obtaining the identification result of the user, the method further includes:

[0062] If the identification result of the user indicates that the user is a low-age Internet complaint user, manual intervention is performed on the user.

[0063] Specifically, if the user is identified as a low-Internet-age complaint user based on the identification results, the complaint will be manually intervened. For example, a telephone follow-up will be made to the user, or targeted text messages will be pushed, door-to-door care, free care phone bills and other measures will be taken to ensure that the user will not make complaints within the preset time period.

[0064] It is worth mentioning that the present invention provides a method for identifying low-network age complaint users based on neural networks. By inputting various parameters of new network users (network access data and service data), it is used to identify potential complaint users so that operators can intervene in advance to prevent such users from causing losses to operators and curb such situations from occurring at the source.

[0065] In a specific embodiment, a method for identifying users with low Internet age complaints based on a neural network is provided.

[0066] It should be noted that the definition of low-age users in business indicators is that users with a difference of less than 6 months from account opening to account closing are considered low-age users. Users who make complaints within these 6 months are considered low-age users. Because such users bring no substantial income to the company and require the company to pay additional costs, they cause considerable trouble to the company, especially the customer service department. If the model can be used to identify them in advance, early intervention can be made to reduce the company's losses.

[0067] The present invention collects the information of user access to the network (i.e., access data) as basic data, and also includes some business data. Business data refers to the business type data generated by users when using the operator's business. Among them, the data of user access information mainly include access time and complaint time (only required for the training process), and the business data mainly include development channels, account opening categories, product identification, whether to access the network again (the above indicators are indicators that the model in the access data and business data can directly use for identification), etc., and are input into the training model to obtain the recognition result at the same time, and the model is compared by accuracy, so as to fix an accurate and efficient recognition method. Access time, complaint time, development channel, account opening category, product identification, whether to access the network again, these indicators are represented by a, b, c, d, e, and f respectively. The problem can be equivalent to predicting the time t of the user's first complaint, and superimposing the access to the network for less than 6 months, that is, for example, if a user generates a complaint in the 5th month after accessing the network, then it is a complaint from a user with low network age, and this type of user can be used as the identification object of the present invention. If the user complained in the 8th month after joining the network, the user has been online for more than 6 months, which is the definition of a user with low Internet age. Then the user does not belong to the target user for identification. A method for identifying complaints from users with low Internet age can be obtained.

[0068] To use a neural network to predict the time t when a user makes their first complaint, we need to transform the problem into a supervised learning problem, where a, b, c, d, e, and f are input features, t is the target variable to be predicted, and a neural network is used for training and reasoning.

[0069] Specifically, during the training process, data preprocessing is required to encode category features. One-hot encoding can be used to process development channels, account opening categories, product identification, and whether to re-enter the network. Convert it into a binary encoding form for machine recognition. Then use normalization to process the network access time and complaint time to improve training efficiency. Define the input layer to build a neural network model. The input dimension is the number of features. Add one or more hidden layers. Each hidden layer has a set of weights w and biases b. Define the output layer to have only one dimension t, which is the value to be predicted. After defining the neural network model, you need to initialize the network parameters and give the weight w and bias b an initial random value respectively. Then calculate the output z of each layer and perform forward propagation. The calculation formula is:

[0070] z i =f(z i )

[0071] Where i represents the layer index and f(x) is the activation function. The following functions can be used:

[0072]

[0073] Next, calculate the loss L, and use the mean square error as the loss function, that is:

[0074]

[0075] Where N is the sample size, t i is the actual complaint time of the i-th sample, is the predicted complaint time of the i-th sample, and then back propagation is calculated to calculate the gradient of the loss function with respect to each parameter:

[0076]

[0077] Next, update the parameters using the gradient descent method:

[0078]

[0079]

[0080] Among them, α is the learning rate, and the last step is iterative optimization. The process from forward propagation to parameter update can be repeated until the accuracy of the model in the validation set no longer improves, and then this process can be stopped.

[0081] Specifically, the inference process is to load the training model, obtain the weights w and bias b through learning, and simultaneously perform data preprocessing. If all data sets are preprocessed during the training process, this step can be ignored, and then calculate the forward propagation to obtain the predicted output:

[0082] t δ =f(w i +b i )

[0083] Among them, t δ is the final output index. The predicted output t is a continuous value. At this time, the complaint time can be used for analysis. Specifically, the predicted t is obtained through this process. δ , combined with the data of users who have been online for less than 6 months, users with low Internet age and who have complained about the service can be identified for further analysis and decision-making.

[0084] It should be noted that there is an existing deep learning framework for the above whole process, and the present invention is only an example of its implementation process. The existing deep learning framework is PYTorch, which has provided a ready-made high-level API for the whole process from construction to training to reasoning.

[0085] In another specific embodiment, the specific application steps of the method for identifying users with low Internet age complaints mainly include:

[0086] 1. Collect business-side data of all users in the large network. User data directly includes network access data and business data. Network access data includes but is not limited to the user's network access time, complaint time, etc. The unprocessed business data generated by the system side includes development channels, account opening categories, product identification, whether to re-access the network, service usage, historical complaint records, real-time bills and other business-related data.

[0087] Clean the data and handle missing values, outliers, and duplicate records.

[0088] Encode categorical data, such as One-Hot Encoding.

[0089] Normalize or standardize continuous data.

[0090] 2. Based on business knowledge, select key features related to the complaint behavior, such as network access time, service usage frequency, etc. This step screens the selected business data and only selects data related to the complaint behavior.

[0091] Construct new features, such as user activity and service failure rate, where user activity refers to the number of historical complaints. Less than 3 times indicates inactivity, more than 3 times but less than 10 times indicates average, and more than 10 times indicates activity. This helps better capture user complaint tendencies.

[0092] Perform feature selection or dimensionality reduction to reduce model complexity and overfitting risk.

[0093] 3. Label user data to determine which users are low-age Internet users.

[0094] Specifically, create labels such as "complaint" or "no complaint" in order to train a classification model.

[0095] 4. Design a neural network architecture that is appropriate for the problem, such as a multilayer perceptron, convolutional neural network, or recurrent neural network, depending on the data type.

[0096] Determine the number of network layers, the number of neurons in each layer, activation functions and other parameters.

[0097] 5. Use the labeled dataset to train the neural network model.

[0098] Choose an appropriate loss function, such as cross entropy loss, for binary classification problems.

[0099] Choose an optimization algorithm, such as Adam or SGD, to update the network weights.

[0100] 6. Use the validation set to evaluate the performance of the model, such as accuracy, recall, F1 score, etc.

[0101] Adjust model parameters and network structure and perform multiple iterations to improve model performance.

[0102] Use techniques such as Early Stopping and Regularization to prevent overfitting.

[0103] 7. Analyze the importance of each feature in the model and determine which features are most critical for identifying low-age Internet users.

[0104] 8. Deploy the trained model to the production environment.

[0105] Specifically, an API interface is developed to enable the business system to call the model in real time to conduct user complaint risk assessment.

[0106] 9. Monitor the performance of the model in actual applications to ensure its stability and accuracy.

[0107] Regularly retrain the model with new data (update the model) to adapt to changes in user behavior.

[0108] 10. Apply the prediction results of the model to customer service and product improvement, such as targeted customer care, service optimization, etc. If a customer is identified as a low-age complaining user, the complaint will be manually intervened. This part of the user will be called back, or targeted SMS push, door-to-door care, free care phone bills and other measures will be taken to ensure that the user will not make complaints within 6 months.

[0109] It is worth mentioning that the present invention uses a neural network to identify users with low network age complaints based on user network access data and service data. The present invention uses a neural network to predict the target variable t, and combines the network access time to obtain users with low network age complaints. The present invention uses the gradient descent method to train network access data and service data to tend to the accuracy of result prediction. The present invention iteratively optimizes the training process to obtain learning weights and biases, and further performs reasoning to reduce data overfitting.

[0110] The method for identifying users with low network age complaints provided by an embodiment of the present invention first obtains the user's network access data and service data; then inputs the user's network access data and service data into a trained neural network model, and uses the trained neural network model to identify whether the user is a user with low network age complaints, thereby obtaining an identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users. The present invention uses a trained neural network model to identify users with low network age complaints based on the user's network access data and service data, so that operators can intervene in advance to prevent users with low network age complaints from causing losses to operators, thereby curbing such situations from occurring at the source, and solving the problem that the prior art lacks a solution for identifying users with low network age complaints.

[0111] Embodiment 2:

[0112] like Figure 2 As shown, this embodiment provides a device for identifying users with low Internet age complaints, which is used to execute the above-mentioned method for identifying users with low Internet age complaints, including:

[0113] The acquisition module 11 is used to acquire the user's network access data and service data;

[0114] The identification module 12 is connected to the acquisition module 11, and is used to input the user's network access data and service data into a trained neural network model, identify whether the user is a low-network age complaint user through the trained neural network model, and obtain the identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users.

[0115] Furthermore, the device also includes:

[0116] An obtaining module, used for obtaining a training data set according to the network access data and service data of the plurality of historical users;

[0117] A labeling module, used to label whether each historical user in the training data set is a low-age Internet complaint user, to obtain a labeled training data set;

[0118] The training module is used to train the pre-selected neural network model using the labeled training data set to obtain the trained neural network model.

[0119] Furthermore, the obtaining module specifically includes:

[0120] A selection and construction unit, configured to select key features related to the complaint behavior in the business data of the multiple historical users according to the business data and business knowledge of the multiple historical users, and construct new features of the multiple historical users;

[0121] As a unit, it is used to use the network access data, service data, key features related to complaint behavior and new features of the multiple historical users as the training data set.

[0122] Furthermore, the neural network model is one of the following: a multi-layer perceptron MLP, a convolutional neural network CNN, or a recurrent neural network RNN.

[0123] Furthermore, the network access data includes: network access time, and the service data includes: development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills;

[0124] The identification module 12 specifically includes:

[0125] The prediction and identification unit is used to input the user's network access time, development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills into the trained neural network model, and obtain the user's key features related to the complaint behavior and the user's new features based on the user's service usage, historical complaint records, and real-time bills through the neural network model, and predict the user's first complaint time based on the user's development channel, account opening category, product identification, whether to re-access the network, and the key features and new features related to the complaint behavior, and identify whether the user is a low-age complaint user based on the user's first complaint time and the user's network access time.

[0126] Furthermore, the prediction and recognition unit includes:

[0127] The identification unit is used to identify the user as a low-network-age complaint user if the difference between the time of the user's first complaint and the time of the user's network access is less than a preset time period.

[0128] Furthermore, the device also includes:

[0129] The intervention module is used to perform manual intervention on the user if the identification result of the user indicates that the user is a user with low Internet age complaint.

[0130] Embodiment 3:

[0131] refer to Figure 3 This embodiment provides a device for identifying users with low Internet age complaint, including a memory 21 and a processor 22. The memory 21 stores a computer program, and the processor 22 is configured to run the computer program to execute the method for identifying users with low Internet age complaint in Example 1.

[0132] The memory 21 is connected to the processor 22. The memory 21 may be a flash memory, a read-only memory or other memory. The processor 22 may be a central processing unit or a single-chip microcomputer.

[0133] Embodiment 4:

[0134] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method for identifying users with low Internet age complaints in the above-mentioned embodiment 1 is implemented.

[0135] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable read only memory), flash memory or other memory technology, CD-ROM (Compact Disc Read-Only Memory), digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer.

[0136] In summary, the method, device and readable storage medium for identifying users with low network age complaints provided by the embodiments of the present invention first obtain the user's network access data and service data; then input the user's network access data and service data into a trained neural network model, and use the trained neural network model to identify whether the user is a user with low network age complaints, and obtain the identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users. The present invention uses a trained neural network model to identify users with low network age complaints based on the user's network access data and service data, so that operators can intervene in advance to prevent users with low network age complaints from causing losses to operators, thereby curbing such situations from happening at the source, and solving the problem that the prior art has no relevant solutions for how to identify users with low network age complaints.

[0137] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for identifying users with low Internet age complaints, characterized in that: The method comprises: Obtain the user's network access data and business data; The network access data and service data of the user are input into a trained neural network model, and the trained neural network model is used to identify whether the user is a low-network age complaint user to obtain an identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users.

2. The method according to claim 1, characterized in that Before inputting the network access data and service data of the user into the trained neural network model, identifying whether the user is a low-age network complaint user through the trained neural network model, and obtaining the identification result of the user, the method further includes: Obtaining a training data set according to the network access data and service data of the plurality of historical users; Marking whether each historical user in the training data set is a low-age Internet complaint user to obtain a marked training data set; The pre-selected neural network model is trained using the labeled training data set to obtain the trained neural network model.

3. The method according to claim 2, characterized in that The obtaining of a training data set according to the network access data and service data of the plurality of historical users specifically includes: According to the business data and business knowledge of the multiple historical users, key features related to the complaint behavior in the business data of the multiple historical users are selected, and new features of the multiple historical users are constructed; The network access data, service data, key features related to the complaint behavior and new features of the multiple historical users are used as the training data set.

4. The method according to claim 1, characterized in that: The neural network model is one of the following: multi-layer perceptron MLP, convolutional neural network CNN, recurrent neural network RNN.

5. The method according to claim 1, characterized in that The network access data includes: network access time, and the business data includes: development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills; The inputting the network access data and service data of the user into the trained neural network model, identifying whether the user is a low-network age complaint user through the trained neural network model, and obtaining the identification result of the user specifically includes: The user's network access time, development channel, account opening category, product identification, whether to re-access the network, service usage, historical complaint records, and real-time bills are input into the trained neural network model, and the neural network model obtains the user's key features related to the complaint behavior and the user's new features based on the user's service usage, historical complaint records, and real-time bills. The user's first complaint time is predicted based on the user's development channel, account opening category, product identification, whether to re-access the network, the key features related to the complaint behavior, and the new features. Based on the user's first complaint time and the user's network access time, it is identified whether the user is a low-network-age complaint user.

6. The method according to claim 5, characterized in that The identifying, according to the time of the first complaint of the user and the time of the user joining the network, whether the user is a user with low network age making complaints specifically includes: If the difference between the time of the user's first complaint and the time of the user's network access is less than a preset time period, the user is identified as a low-network-age complaint user.

7. The method according to claim 1, characterized in that After inputting the network access data and service data of the user into the trained neural network model, identifying whether the user is a low-age network complaint user through the trained neural network model, and obtaining the identification result of the user, the method further includes: If the identification result of the user indicates that the user is a low-age Internet complaint user, manual intervention is performed on the user.

8. A device for identifying users with low Internet age complaints, characterized in that: include: The acquisition module is used to obtain the user's network access data and service data; An identification module is connected to the acquisition module and is used to input the user's network access data and service data into a trained neural network model, identify whether the user is a low-network age complaint user through the trained neural network model, and obtain an identification result of the user, wherein the trained neural network model is trained based on the network access data and service data of multiple historical users.

9. A device for identifying users with low Internet age complaints, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the method for identifying users with low Internet age complaints as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the method for identifying users with low Internet age complaints as described in any one of claims 1 to 7 is implemented.

Images