Derogatory user identification method and device, equipment, storage medium and product

Abstract

The invention discloses a derogatory user identification method and device, equipment, a storage medium and a product, and the method comprises the steps: obtaining evaluation data of a user for a mobile network, and removing malicious evaluation data in the evaluation data, thereby obtaining first evaluation data; wherein the evaluation data comprises a terminal type of a user, a user identifier, an evaluation opinion and an evaluation result; performing operations of word segmentation, stop word removal and word bag division on the first evaluation data to obtain second evaluation data; and adopting a preset naive Bayesian model to identify the second evaluation data, and outputting derogatory users of the mobile network. Malicious evaluation data of a user to a mobile network can be filtered, a customer evaluation prediction model is established by using a NLP sentiment analysis method of naive Bayes, whether the user has a derogatory sentiment tendency or not is analyzed, the recognition accuracy of derogatory users is improved, the derogatory users are repaired, and then the user satisfaction degree is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of network operators, and in particular to a method, device, equipment, storage medium and product for identifying derogatory users. Background Art

[0002] Existing technologies generally use questionnaires to collect user evaluation information on mobile operators' networks, products and services, evaluate network quality, understand user experience and feelings, and measure user satisfaction with mobile networks. By locating the time and place when users are dissatisfied, correlating the O / B domain data (operation / business domain) of the mobile network, finding out the user's relevant information in terms of network, package, tariff, etc., analyzing the reasons that may lead to dissatisfaction or complaints, formulating plans, and repairing the devalued users, user satisfaction is improved;

[0003] However, the user evaluation information obtained through survey questionnaires is generally difficult to reflect the issues that users are concerned about and their actual situations. In addition, small sample survey questionnaires may cause sampling bias. Therefore, the authenticity of user satisfaction with mobile networks obtained through survey questionnaires cannot be guaranteed. It is also difficult to accurately identify derogatory users of mobile networks, making it difficult to accurately repair derogatory users. Summary of the invention

[0004] The present invention provides a method, device, equipment, storage medium and product for identifying derogatory users. The method, based on filtering malicious evaluation data of users on mobile networks, uses the NLP sentiment analysis method of naive Bayes to establish a customer evaluation prediction model, analyzes whether the user has a derogatory emotional tendency, improves the accuracy of identifying derogatory users, repairs derogatory users, and thereby improves user satisfaction.

[0005] In order to achieve the above object, an embodiment of the present invention provides a method for identifying derogatory users, comprising:

[0006] Acquire the user's evaluation data on the mobile network, remove malicious evaluation data from the evaluation data, and obtain first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion and evaluation result;

[0007] Perform word segmentation, stop word removal, and word bag division operations on the first evaluation data to obtain second evaluation data;

[0008] The second evaluation data is identified using a preset naive Bayes model, and derogatory users of the mobile network are output.

[0009] As an improvement of the above solution, the step of obtaining the user's evaluation data on the mobile network and removing malicious evaluation data from the evaluation data to obtain the first evaluation data includes:

[0010] Obtaining user evaluation data on the mobile network, and entering the evaluation data into a user emotion recognition library after normalizing the evaluation data;

[0011] If the normalized evaluation data is malicious, the malicious evaluation data will be used as the malicious review training set, and the other evaluation data will be used as the malicious review test set;

[0012] Constructing feature vectors of the malicious comment training set and the malicious comment test set;

[0013] Malicious evaluation data in the malicious comment test set is eliminated according to the feature vector to obtain first evaluation data.

[0014] As an improvement of the above solution, the step of using a preset naive Bayes model to identify the second evaluation data and outputting derogatory users of the mobile network includes:

[0015] Generate the objective functions of positive and negative reviews based on the preset naive Bayes model;

[0016] Identify the second evaluation data according to the objective function of the good and bad reviews to obtain a good review value and a bad review value for each second evaluation data;

[0017] If the favorable review value is less than the unfavorable review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network.

[0018] As an improvement of the above scheme, the objective function of the positive review is:

[0019]

[0020] The objective function of the negative review is:

[0021]

[0022] In the formula, P(D i |G) is the second evaluation data of the ith item, which is a positive review. i is the i-th second evaluation data, P(G) is the probability that the second evaluation data is a good review; N j is the frequency of each word after the comment data is segmented; N is the sum of the word frequencies in the word bag; m is the number of features in the word bag; P(D i |B) is the probability that the i-th second evaluation data is a bad review, and P(B) is the probability that the second evaluation data is a bad review.

[0023] In order to achieve the above-mentioned object, an embodiment of the present invention provides a derogatory user identification device, comprising:

[0024] An evaluation data acquisition module, used to acquire the user's evaluation data on the mobile network, remove malicious evaluation data from the evaluation data, and obtain first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion and evaluation result;

[0025] An evaluation data processing module, used for performing word segmentation, stop word removal and word bag division operations on the first evaluation data to obtain second evaluation data;

[0026] The derogatory user identification module is used to identify the second evaluation data using a preset naive Bayes model and output the derogatory users of the mobile network.

[0027] As an improvement of the above solution, the evaluation data acquisition module is used to:

[0028] Obtaining user evaluation data on the mobile network, and entering the evaluation data into a user emotion recognition library after normalizing the evaluation data;

[0029] If the normalized evaluation data is malicious, the malicious evaluation data will be used as the malicious review training set, and the other evaluation data will be used as the malicious review test set;

[0030] Constructing feature vectors of the malicious comment training set and the malicious comment test set;

[0031] Malicious evaluation data in the malicious comment test set is eliminated according to the feature vector to obtain first evaluation data.

[0032] As an improvement of the above solution, the derogatory user identification module is used to:

[0033] Generate the objective functions of positive and negative reviews based on the preset naive Bayes model;

[0034] Identify the second evaluation data according to the objective function of the good and bad reviews to obtain a good review value and a bad review value for each second evaluation data;

[0035] If the favorable review value is less than the unfavorable review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network.

[0036] In order to achieve the above-mentioned purpose, an embodiment of the present invention provides a derogatory user identification device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the above-mentioned derogatory user identification method when executing the computer program.

[0037] In order to achieve the above-mentioned purpose, an embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned derogatory user identification method.

[0038] To achieve the above objective, an embodiment of the present invention further provides a computer program product, which is stored in a storage medium and is executed by at least one processor to implement the steps of the above derogatory user identification method.

[0039] Compared with the prior art, the embodiment of the present invention discloses a derogatory user identification method, device, equipment, storage medium and product, which obtains the user's evaluation data on the mobile network, removes the malicious evaluation data in the evaluation data, and obtains the first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion and evaluation result; performs word segmentation, stop word removal and word bag division operations on the first evaluation data to obtain the second evaluation data; uses a preset naive Bayes model to identify the second evaluation data and outputs the derogatory users of the mobile network. It can filter the user's malicious evaluation data on the mobile network, use the naive Bayes NLP sentiment analysis method to establish a customer evaluation prediction model, analyze whether the user has a derogatory emotional tendency, improve the recognition accuracy of derogatory users, repair the derogatory users, and thus improve user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a flow chart of a method for identifying derogatory users provided by an embodiment of the present invention;

[0041] Figure 2 is a schematic structural diagram of a derogatory user identification device provided by an embodiment of the present invention;

[0042] Figure 3 The present invention is a structural block diagram of a derogatory user identification device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 creative work are within the scope of protection of the present invention.

[0044] It should be noted that the terms "comprises" and "specifically" and any variations of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0045] See also Figure 1 , Figure 1 1 is a flow chart of a method for identifying a derogatory user provided by an embodiment of the present invention. The method for identifying a derogatory user includes:

[0046] S1, obtaining user evaluation data on a mobile network, removing malicious evaluation data from the evaluation data, and obtaining first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion, and evaluation result;

[0047] S2, performing word segmentation, stop word removal and word bag division operations on the first evaluation data to obtain second evaluation data;

[0048] S3, using a preset naive Bayes model to identify the second evaluation data, and outputting derogatory users of the mobile network.

[0049] Exemplarily, the embodiment of the present invention collects user evaluation data, cleans and processes the data, and uses the Naive Bayes model and NLP sentiment analysis method to predict whether the user has a derogatory tendency, thereby laying the foundation for user repair and satisfaction improvement.

[0050] Exemplarily, in step S2, Chinese word segmentation technology is used to perform word segmentation on user evaluation opinions to obtain user evaluation data after word segmentation; stop words are removed from the user evaluation data after word segmentation; and bag-of-words division is performed on the user evaluation data after stop words. Stop words refer to words and punctuation marks that are meaningless or not very meaningful to user sentiment analysis. Stop words include: punctuation marks, function words, high-frequency but non-emotional words, etc. High-frequency but non-emotional words include: "feel", "mobile phone", "this year", etc. Stop words are stored in a stop word library. The bag of words implements word frequency statistics for different types of user evaluation data, which are divided into: Positive BoW and Negative BoW. Positive BoW stores word frequency statistics for user evaluation results that are satisfactory; Negative BoW stores word frequency statistics for user evaluation results that are unsatisfactory.

[0051] Specifically, the obtaining of user evaluation data on the mobile network, removing malicious evaluation data from the evaluation data, and obtaining first evaluation data includes:

[0052] Obtaining user evaluation data on the mobile network, and entering the evaluation data into a user emotion recognition library after normalizing the evaluation data;

[0053] If the normalized evaluation data is malicious, the malicious evaluation data will be used as the malicious review training set, and the other evaluation data will be used as the malicious review test set;

[0054] Constructing feature vectors of the malicious comment training set and the malicious comment test set;

[0055] Malicious evaluation data in the malicious comment test set is eliminated according to the feature vector to obtain first evaluation data.

[0056] Exemplarily, in order to ensure the objectivity of the evaluation information and the accuracy of the model, the K nearest neighbor algorithm is used to filter malicious comment data before modeling. For example, the user data center of a mobile operator collects a large amount of user evaluation information, mainly including fields such as user identification, evaluation opinions, and evaluation results. Normalize the user evaluation information and enter the processed data into the user emotion recognition library. In the user emotion recognition library, analyze a portion of the user evaluation data to identify whether the evaluation data is a malicious evaluation. If so, use the malicious evaluation data as a malicious comment training set, and other evaluation data as a malicious comment test set; perform word segmentation on each evaluation in the malicious comment training set, store all the keywords after word segmentation in the feature library, and construct a feature vector for the malicious comment training set:

[0057]

[0058] Among them, FV train_i represents the m evaluation data of the malicious comment training set, [a i1 ,a i2 ,...,a ik ] represents the feature vector of the i-th comment in the malicious comment training set, k refers to the number of words in the feature library, if a word in the comment appears in the feature library, the corresponding a is 1; otherwise it is 0.

[0059] Construct the feature vector of the malicious comment test set:

[0060]

[0061] Among them, FV test_i represents n reviews of the malicious comment test set, [b i1 ,b i2 ,...,b ik ] represents the feature vector of the i-th comment in the malicious comment test set, k refers to the number of words in the feature library, if a word in the comment appears in the feature library, the corresponding b is 1; otherwise it is 0.

[0062] Iterate over each comment in the test set and use the Euclidean distance to calculate the distance between a comment in the test set and each comment in the training set.

[0063]

[0064] Output distance matrix:

[0065]

[0066] Among them, a k represents the kth feature vector in the malicious comment training set, b k represents the kth feature vector in the malicious comment test set, d ij Represents the distance between the i-th comment information in the test set and the j-th comment information in the training set.

[0067] In the distance matrix, calculate the 3 (K=3) nearest distances [d1, d2, d3] from each evaluation information in the test set to the training set, analyze the corresponding 3 neighbors [an1, an2, an3], and determine the neighbor's label (whether it is a malicious comment). If the number of neighbors with malicious labels is greater than the number of neighbors with non-malicious labels, then the evaluation is a malicious comment and needs to be filtered. After filtering all malicious evaluation data in the malicious comment test set, the first evaluation data is obtained.

[0068] Specifically, the adopting a preset naive Bayes model to identify the second evaluation data and outputting the derogatory users of the mobile network includes:

[0069] Generate the objective functions of positive and negative reviews based on the preset naive Bayes model;

[0070] Identify the second evaluation data according to the objective function of the good and bad reviews to obtain a good review value and a bad review value for each second evaluation data;

[0071] If the favorable review value is less than the unfavorable review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network.

[0072] Specifically, the objective function of the favorable comments is:

[0073]

[0074] The objective function of the negative review is:

[0075]

[0076] In the formula, P(D i |G) is the second evaluation data of the ith item, which is a positive review. iis the i-th second evaluation data, P(G) is the probability that the second evaluation data is a good review; N j is the frequency of each word after the comment data is segmented; N is the sum of the word frequencies in the word bag; m is the number of features in the word bag; P(D i |B) is the probability that the i-th second evaluation data is a bad review, and P(B) is the probability that the second evaluation data is a bad review.

[0077] For example, in the user emotion recognition library, in addition to satisfaction and dissatisfaction, there are also blank information in the comment results. Using NLP sentiment analysis method, based on the naive Bayes model, it is predicted whether the user who has not commented belongs to the derogatory customer group. The main steps of the sentiment prediction module are as follows:

[0078] Filter user comments that do not provide satisfaction information from the user emotion recognition library center and obtain comment opinions D i (i=1,2,3,...);

[0079] Traverse D i (i=1,2,3,...), such as: "My family installed mobile broadband last year, the speed is OK, I can watch videos and play games without lag..."; for D i Evaluation text after word segmentation and processing D′ j For each word after the word segmentation. For example: D i =|My family installed a mobile broadband last year, the speed is OK, I can watch videos and play games without any lags..., D′ j =[|my home|,|last year|,……,|not stuck|];

[0080] Call the stop word list to D i Filter the stop word list and remove the stop words Remove useless words and keep words that are useful for identifying derogatory users. An example of the review text after filtering the stop word list is as follows: D i =|Internet speed|is|ok|for|watching|videos|and|playing|games|without|lag|;

[0081] Calculate whether the review information is a positive or negative review, a positive review is P(G|D i ), and the negative rating is P(B|D i ), according to the Naive Bayes model: but: Since we only need to compare the size and the denominators are the same, we only need to calculate the numerator. The objective function of the positive review is: P(D i |G)*P(G), the objective function of the negative review is: P(D i |B)*P(B).

[0082] in, D i The frequency of each word in the word bag (Positive Bow) after word segmentation, N is the sum of the word frequencies in the word bag (Positive Bow), Among them, count(Positive Bow) represents the number of records of satisfactory comments, and count(Negative Bow) represents the number of records of unsatisfactory comments.

[0083] in, D i The frequency of each word in the negative bow after word segmentation, N is the sum of the frequency of the words in the negative bow,

[0084]

[0085] The number of records of comments, count(Negative Bow) indicates the number of records of dissatisfied comments.

[0086] Considering that when the frequency of a word is 0, the probability will be 0, so the Laplace smoothing coefficient is introduced Optimize. i It is the frequency of each word after the comment information is segmented; α can be customized and is 1 by default; N is the sum of the word frequencies in the word bag; m is the number of features in the word bag.

[0087] The objective function of the positive and negative reviews is recalculated based on the Laplace smoothing coefficient, and if the positive review value is less than the negative review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network. The above process is repeated to finally output all derogatory users.

[0088] The embodiment of the present invention discloses a method for identifying derogatory users, which obtains the evaluation data of users on a mobile network, removes malicious evaluation data from the evaluation data, and obtains first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion, and evaluation result; performs word segmentation, stop word removal, and word bag division operations on the first evaluation data to obtain second evaluation data; uses a preset naive Bayes model to identify the second evaluation data, and outputs derogatory users of the mobile network. The method can filter out malicious evaluation data of users on a mobile network, use the naive Bayes NLP sentiment analysis method to establish a customer evaluation prediction model, analyze whether users have derogatory sentiment tendencies, improve the accuracy of identifying derogatory users, repair derogatory users, and thus improve user satisfaction.

[0089] See also Figure 2 , Figure 2 is a schematic diagram of the structure of a derogatory user identification device 10 provided in an embodiment of the present invention. The derogatory user identification device 10 comprises:

[0090] The evaluation data acquisition module 11 is used to obtain the user's evaluation data on the mobile network, remove malicious evaluation data from the evaluation data, and obtain first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion and evaluation result;

[0091] An evaluation data processing module 12, configured to perform word segmentation, stop word removal and word bag division operations on the first evaluation data to obtain second evaluation data;

[0092] The derogatory user identification module 13 is used to identify the second evaluation data using a preset naive Bayes model and output the derogatory users of the mobile network.

[0093] Specifically, the evaluation data acquisition module 11 is used to:

[0094] Obtaining user evaluation data on the mobile network, and entering the evaluation data into a user emotion recognition library after normalizing the evaluation data;

[0095] If the normalized evaluation data is malicious, the malicious evaluation data will be used as the malicious review training set, and the other evaluation data will be used as the malicious review test set;

[0096] Constructing feature vectors of the malicious comment training set and the malicious comment test set;

[0097] Malicious evaluation data in the malicious comment test set is eliminated according to the feature vector to obtain first evaluation data.

[0098] Specifically, the derogatory user identification module 13 is used to:

[0099] Generate the objective functions of positive and negative reviews based on the preset naive Bayes model;

[0100] Identify the second evaluation data according to the objective function of the good and bad reviews to obtain a good review value and a bad review value for each second evaluation data;

[0101] If the favorable review value is less than the unfavorable review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network.

[0102] A derogatory user identification device 10 provided in an embodiment of the present invention can implement all the processes of the derogatory user identification method of the above embodiment. The functions of each module in the device and the technical effects achieved are respectively the same as the functions and technical effects achieved by the derogatory user identification method of the above embodiment, and will not be repeated here.

[0103] See also Figure 3 , Figure 3 1 is a schematic diagram of the structure of a derogatory user identification device 20 provided in an embodiment of the present invention. The derogatory user identification device 20 of this embodiment includes: a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. When the processor 21 executes the computer program, the steps in the above-mentioned derogatory user identification method embodiment are implemented. Alternatively, when the processor 21 executes the computer program, the functions of each module in the above-mentioned derogatory user identification device embodiment are implemented.

[0104] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory 22 and executed by the processor 21 to implement the present invention. The one or more modules may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the derogatory user identification device 20.

[0105] The derogatory user identification device 20 may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The derogatory user identification device 20 may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art may understand that the schematic diagram is only an example of the derogatory user identification device 20 and does not constitute a limitation on the derogatory user identification device 20. The derogatory user identification device 20 may include more or less components than shown in the figure, or combine certain components, or different components. For example, the derogatory user identification device 20 may also include input and output devices, network access devices, buses, etc.

[0106] The processor 21 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 21 is the control center of the derogatory user identification device 20, and uses various interfaces and lines to connect various parts of the entire derogatory user identification device 20.

[0107] The memory 22 can be used to store the computer program and / or module. The processor 21 realizes various functions of the derogatory user identification device 20 by running or executing the computer program and / or module stored in the memory 22 and calling the data stored in the memory 22. The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0108] Wherein, if the module integrated in the derogatory user identification device 20 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor 21, the steps of the above-mentioned method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0109] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0110] An embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the derogatory user identification method as described in the above embodiment.

[0111] In addition, an embodiment of the present invention further provides a computer program product, which is stored in a storage medium and is executed by at least one processor to implement the steps of the derogatory user identification method of the above embodiment.

[0112] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for identifying derogatory users, characterized in that: include: Acquire the user's evaluation data on the mobile network, remove malicious evaluation data from the evaluation data, and obtain first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion and evaluation result; Perform word segmentation, stop word removal, and word bag division operations on the first evaluation data to obtain second evaluation data; The second evaluation data is identified using a preset naive Bayes model, and derogatory users of the mobile network are output.

2. The method for identifying derogatory users according to claim 1, characterized in that: The step of obtaining the user's evaluation data on the mobile network and removing malicious evaluation data from the evaluation data to obtain first evaluation data includes: Obtaining user evaluation data on the mobile network, and entering the evaluation data into a user emotion recognition library after normalizing the evaluation data; If the normalized evaluation data is malicious, the malicious evaluation data will be used as the malicious review training set, and the other evaluation data will be used as the malicious review test set; Constructing feature vectors of the malicious comment training set and the malicious comment test set; Malicious evaluation data in the malicious comment test set is eliminated according to the feature vector to obtain first evaluation data.

3. The method for identifying derogatory users according to claim 1, characterized in that: The adopting a preset naive Bayes model to identify the second evaluation data and outputting the derogatory users of the mobile network includes: Generate the objective functions of positive and negative reviews based on the preset naive Bayes model; Identify the second evaluation data according to the objective function of the good and bad reviews to obtain a good review value and a bad review value for each second evaluation data; If the favorable review value is less than the unfavorable review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network.

4. The method for identifying derogatory users according to claim 1, characterized in that: The objective function of the positive review is: The objective function of the negative review is: In the formula, P(D i |G) is the second evaluation data of the ith item, which is a positive review. i is the i-th second evaluation data, P(G) is the probability that the second evaluation data is a good review; N j is the frequency of each word after the comment data is segmented; N is the sum of the word frequencies in the word bag; m is the number of features in the word bag; P(D i |B) is the probability that the i-th second evaluation data is a bad review, and P(B) is the probability that the second evaluation data is a bad review.

5. A derogatory user identification device, characterized in that include: An evaluation data acquisition module, used to acquire the user's evaluation data on the mobile network, remove malicious evaluation data from the evaluation data, and obtain first evaluation data; wherein the evaluation data includes the user's terminal type, user identification, evaluation opinion and evaluation result; An evaluation data processing module, used for performing word segmentation, stop word removal and word bag division operations on the first evaluation data to obtain second evaluation data; The derogatory user identification module is used to identify the second evaluation data using a preset naive Bayes model and output the derogatory users of the mobile network.

6. The derogatory user identification device according to claim 5, characterized in that: The evaluation data acquisition module is used to: Obtaining user evaluation data on the mobile network, and entering the evaluation data into a user emotion recognition library after normalizing the evaluation data; If the normalized evaluation data is malicious, the malicious evaluation data will be used as the malicious review training set, and the other evaluation data will be used as the malicious review test set; Constructing feature vectors of the malicious comment training set and the malicious comment test set; Malicious evaluation data in the malicious comment test set is eliminated according to the feature vector to obtain first evaluation data.

7. The derogatory user identification device according to claim 5, characterized in that: The derogatory user identification module is used to: Generate the objective functions of positive and negative reviews based on the preset naive Bayes model; Identify the second evaluation data according to the objective function of the good and bad reviews to obtain a good review value and a bad review value for each second evaluation data; If the favorable review value is less than the unfavorable review value, the user corresponding to the corresponding second evaluation data is a derogatory user of the mobile network.

8. A derogatory user identification device, characterized in that The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for identifying derogatory users as claimed in any one of claims 1 to 4 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the derogatory user identification method according to any one of claims 1 to 4.

10. A computer program product, characterized in that The computer program product is stored in a storage medium, and the program product is executed by at least one processor to implement the steps of the derogatory user identification method according to any one of claims 1 to 6.

Images