Data processing method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application disclose a data processing method and device, electronic equipment and storage medium. The method comprises: an electronic device obtains initial feedback samples corresponding to a plurality of survey users, and extracts initial derogatory keywords from the initial feedback samples to obtain an initial derogatory keyword set; similar users similar to a predicted user are determined from the plurality of survey users, and first feedback samples corresponding to the similar users in the initial feedback samples are determined; first derogatory keywords appearing in the initial derogatory keyword set are determined from the first feedback samples; the confidence of each first derogatory keyword in the first feedback samples is determined; and target derogatory keywords corresponding to the predicted user are determined from the first derogatory keywords according to the confidence. Thus, the derogatory reasons of the predicted user are determined according to the surveyed users and the initial feedback samples corresponding to the surveyed users.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, in particular to a data processing method and device, electronic equipment and storage medium. BACKGROUND

[0002] In the current flourishing development of the Internet, a good reputation can bring huge traffic to an enterprise and improve user stickiness. Many Internet companies use the method of obtaining the Net Promoter Score (NPS) to determine the possibility of users willing to recommend products or services to others.

[0003] However, the existing method of investigating the Net Promoter Score is mainly in the form of a questionnaire. This method has a limited coverage and poor timeliness in collecting the questionnaire, and cannot accurately determine the reasons for user depreciation, resulting in the inability to retain users in time. SUMMARY

[0004] The embodiments of the present application provide a data processing method, device, electronic equipment and storage medium. The data processing method can determine the reasons for depreciation of a predicted user according to the initial feedback samples of the investigated users and the initial feedback samples corresponding to the investigated users.

[0005] In a first aspect, the embodiments of the present application provide a data processing method, comprising:

[0006] obtaining initial feedback samples corresponding to a plurality of investigated users, and performing initial depreciation key word extraction on the initial feedback samples to obtain an initial depreciation key word set;

[0007] determining similar users similar to the predicted user from the plurality of investigated users, and determining first feedback samples corresponding to the similar users in the initial feedback samples;

[0008] determining first depreciation key words appearing in the initial depreciation key word set in the first feedback samples;

[0009] determining a confidence degree of each first depreciation key word in the first feedback samples;

[0010] determining a target depreciation key word corresponding to the predicted user from the first depreciation key words according to the confidence degree.

[0011] In a second aspect, the embodiments of the present application provide a data processing device, comprising:

[0012] an obtaining module configured to obtain initial feedback samples corresponding to a plurality of investigated users, and perform initial depreciation key word extraction on the initial feedback samples to obtain an initial depreciation key word set;

[0013] The first determining module is configured to determine similar users similar to the predicted user from the plurality of investigation users, and determine corresponding first feedback samples of the similar users in the initial feedback samples.

[0014] The second determining module is configured to determine first derogatory keywords appearing in the initial derogatory keyword set from the first feedback samples.

[0015] The third determining module is configured to determine a confidence degree of each first derogatory keyword in the first feedback samples.

[0016] The fourth determining module is configured to determine a target derogatory keyword corresponding to the predicted user from the first derogatory keywords according to the confidence degrees.

[0017] In a third aspect, an electronic device is provided, including a memory storing executable program codes, and a processor coupled with the memory; the processor invokes the executable program codes stored in the memory to execute the steps of the data processing method provided in the embodiments of the present application.

[0018] In a fourth aspect, a storage medium is provided, which stores a plurality of instructions, and the instructions are adapted to be loaded by a processor to execute the steps of the data processing method provided in the embodiments of the present application.

[0019] In the embodiments of the present application, the electronic device obtains initial feedback samples corresponding to a plurality of investigation users, and extracts initial derogatory keywords from the initial feedback samples to obtain an initial derogatory keyword set; determines similar users similar to a predicted user from the plurality of investigation users, and determines corresponding first feedback samples of the similar users in the initial feedback samples; determines first derogatory keywords appearing in the initial derogatory keyword set from the first feedback samples; determines a confidence degree of each first derogatory keyword in the first feedback samples; and determines a target derogatory keyword corresponding to the predicted user from the first derogatory keywords according to the confidence degrees. Thus, the derogatory reasons of the predicted user are determined according to the investigation users and the initial feedback samples corresponding to the investigation users. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0021] Figure 1 is a first flowchart of the data processing method provided in the embodiments of the present application.

[0022] Figure 2This is a schematic diagram of the second process of the data processing method provided in the embodiments of this application.

[0023] Figure 3 This is a schematic diagram of the structure of the data processing device provided in the embodiments of this application.

[0024] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] In today's booming internet era, a good reputation can bring huge traffic to businesses and increase user stickiness. Many internet companies use Net Promoter Score (NPS) to determine the likelihood that users are willing to recommend a product or service to others.

[0027] However, existing methods for calculating Net Promoter Score (NPS) primarily rely on questionnaires. These methods have limited reach and poor questionnaire response timeliness, making it difficult to accurately identify the reasons for user dissent and hindering timely user retention.

[0028] To address this technical problem, embodiments of this application provide a data processing method, apparatus, electronic device, and storage medium. This data processing method can determine the reasons for user depreciation based on surveyed users and their corresponding initial feedback samples.

[0029] The data processing method provided in this application can be applied to electronic devices with storage and computing capabilities, such as computers, servers, and workstations.

[0030] Please see Figure 1 , Figure 1 This is a first flowchart illustrating the data processing method provided in an embodiment of this application. The data processing method may include the following steps:

[0031] 110. Obtain initial feedback samples from multiple survey users, and extract initial derogatory keywords from the initial feedback samples to obtain an initial derogatory keyword set.

[0032] In some embodiments, the initial feedback sample of feedback from different survey users can be obtained from online or offline channels through questionnaire survey. For example, the evaluation content of the survey users can be used as the initial feedback sample, or the different options in the questionnaire can be used as the initial feedback sample.

[0033] Specifically, the electronic device can integrate the questionnaire content of the same survey user in the same time period, extract the text information therein, and obtain the initial feedback sample. In the initial feedback sample, there is information of the survey user evaluating and feeding back the product or service.

[0034] It should be noted that one survey user can correspond to multiple initial feedback samples of different times. For example, the survey user fills out a questionnaire in January and fills out another questionnaire in May. The questionnaire in January corresponds to the initial feedback sample of the survey user, and the questionnaire in May corresponds to the initial feedback sample of the survey user.

[0035] In some embodiments, after obtaining the initial feedback sample corresponding to the surveyed user, the initial feedback sample can be subjected to initial derogatory keyword extraction to obtain an initial derogatory keyword set.

[0036] Specifically, the electronic device can perform word segmentation extraction and / or filter stop words on the initial feedback sample to obtain a first processing sample, then input the first processing sample into a text processing model to output initial derogatory keywords and obtain the initial derogatory keyword set.

[0037] For example, the electronic device can first obtain all the text in the initial feedback sample, then perform word segmentation extraction on all the text, such as using a word segmentation extraction tool to extract keywords, which can be various words and short sentences. Thus, the first processing sample is obtained.

[0038] For another example, the electronic device can also set corresponding stop words, such as some positive words, such as “good review”, “not bad”, “good to use”, and other words with positive meaning, which are set as stop words. Then, the stop words are filtered from all the text. Thus, the first processing sample is obtained.

[0039] During the processing of the initial feedback sample, the electronic device can first extract the text of the initial feedback sample to obtain multiple texts corresponding to the initial feedback sample, then filter some words by filtering stop words, and then extract the text after filtering stop words by word segmentation extraction. Thus, the first processing sample is obtained.

[0040] Finally, the first processing sample is processed through the text processing model to obtain the initial derogatory keywords, so as to form an initial derogatory keyword set according to the initial derogatory keywords. For example, the text processing model can obtain the word vector of different keywords in the first processing sample, then generate the initial derogatory keywords according to the word vector, and finally generate the initial derogatory keyword set.

[0041] The text processing model can use a Bert model or a keyBert model to complete keyword extraction of the first processing sample and obtain the initial derogatory keywords. The initial derogatory keywords are words that the research user generates negative evaluations on the product or service.

[0042] 120. Among the plurality of research users, similar users similar to the predicted user are determined, and the similar users are determined to have corresponding first feedback samples in the initial feedback sample.

[0043] Each user has corresponding user characteristics, which can be user age, gender, consumption habits, economic strength, personal preferences, and other multi-dimensional user characteristics.

[0044] In some embodiments, the electronic device can obtain the first feature corresponding to each research user and the second feature of the predicted user, then determine the similarity between each research user and the predicted user according to the first feature and the second feature, and finally determine the research user with a similarity greater than a preset similarity threshold as a similar user.

[0045] For example, the electronic device can determine the user characteristics of each research user according to the personal data corresponding to each research user in the database, such as obtaining the personal data authorized by the user for storage and use in the database, such as age, gender, consumption habits, economic strength, personal preferences, and other multi-dimensional user characteristics. Then generate a corresponding vector for each dimension of the research user characteristics, and finally generate the first feature corresponding to the research user according to the vector of each dimension of the user characteristics, which can be understood as the vector addition of the multi-dimensional research user characteristics.

[0046] The electronic device can determine the user characteristics of each predicted user according to the personal data corresponding to each predicted user in the database, such as obtaining the personal data authorized by the user for storage and use in the database, such as age, gender, consumption habits, economic strength, personal preferences, and other multi-dimensional user characteristics. Then generate a corresponding vector for each dimension of the predicted user characteristics, and finally generate the second feature corresponding to the predicted user according to the vector of each dimension of the user characteristics, which can be understood as the vector addition of the multi-dimensional user characteristics.

[0047] Finally, the electronic device can obtain the feature similarity between each first feature and second feature, such as can adopt a plurality of ways such as cosine distance, Euclidean distance, Pearson correlation coefficient, etc. to obtain the feature similarity between each first feature and second feature.

[0048] The electronic device can determine the feature similarity between each first feature and second feature as the similarity between each investigation user and prediction user, and finally determine the investigation user with a similarity greater than a preset similarity threshold as a similar user.

[0049] In some embodiments, after determining the similar user, the first feedback sample corresponding to the similar user can be determined in the initial feedback sample.

[0050] In some embodiments, the electronic device can calculate the net recommendation value corresponding to each similar user according to the first feedback sample corresponding to each similar user, and then determine the first derogatory user in the similar user according to the net recommendation value.

[0051] For example, the electronic device determines a target similar user with a net recommendation value less than a preset net recommendation value in the similar user, and then determines the target similar user as the first derogatory user.

[0052] In some embodiments, the electronic device can determine the proportion of the first derogatory user in the similar user; if the proportion is greater than a preset proportion threshold, the prediction user is determined as the target derogatory user.

[0053] That is, when the proportion of the first derogatory user in the similar user is greater than the preset proportion threshold, it means that most of the similar users to the prediction user are derogatory users, and the prediction user is also a derogatory user with a high probability.

[0054] 130、Determine the first derogatory keyword appearing in the initial derogatory keyword set in the first feedback sample.

[0055] In some embodiments, the electronic device can also obtain each keyword of the first feedback sample after determining the first feedback sample corresponding to the similar user, and then determine the keyword appearing in the initial derogatory keyword set as the first derogatory keyword.

[0056] For example, the electronic device can match each keyword with the initial derogatory keyword set by keyword matching, and if a keyword matches successfully, the keyword is determined as the first derogatory keyword.

[0057] The electronic device can also determine the number of times each first derogatory keyword appears in the first feedback samples after determining the first derogatory keywords. For example, a first derogatory keyword can be matched with each first feedback sample, and the number of times the first derogatory keyword appears in each first feedback sample can be counted. Finally, the numbers of times are added to obtain the number of times the first derogatory keyword appears in all first feedback samples.

[0058] 140. Determine the confidence of each first derogatory keyword in the first feedback samples.

[0059] In some embodiments, the electronic device can determine a target first derogatory keyword from the first derogatory keywords, determine a first total number of times the target first derogatory keyword appears in all first feedback samples, determine a second total number of times all first derogatory keywords appear in all first feedback samples, and finally divide the first total number by the second total number to obtain the confidence corresponding to the target first derogatory keyword.

[0060] For example, the target first derogatory keyword includes keyword A, keyword B, and keyword C, where keyword A is the target first derogatory keyword. Then, the first total number of times keyword A appears in all first feedback samples is obtained. The electronic device then determines the second total number of times keyword A, keyword B, and keyword C appear in all first feedback samples. Finally, the first total number is divided by the second total number to obtain the confidence corresponding to keyword A.

[0061] In some embodiments, the electronic device can also determine a first similar user corresponding to the target first derogatory keyword, obtain the similarity between each first similar user and the predicted user, add the similarity corresponding to each first similar user to obtain a first sum of similarities, and multiply the first sum of similarities by the first total number of times the target first derogatory keyword appears in all first feedback samples to obtain a first product result.

[0062] The electronic device obtains a second sum of similarities of similar users corresponding to all first derogatory keywords and a second sum of similarities corresponding to each similar user and the predicted user, and then multiplies the second sum of similarities by a second total number of times all first derogatory keywords appear in all first feedback samples to obtain a second product result.

[0063] Finally, the confidence corresponding to the target first derogatory keyword is obtained by dividing the first product result by the second product result.

[0064] In this way, the confidence corresponding to each first keyword in all first feedback samples can be determined in turn.

[0065] 150. Determine the target derogatory keyword corresponding to the predicted user from the first derogatory keyword according to the confidence.

[0066] In some embodiments, the electronic device determines a second derogatory keyword with a confidence greater than a preset confidence from the first derogatory keyword, and determines the second derogatory keyword as the target derogatory keyword corresponding to the predicted user.

[0067] That is, the second derogatory keyword can be used as the derogatory reason of the predicted user, and if the predicted user is a derogatory user, the entire business or product can be adjusted according to the derogatory reason.

[0068] If the predicted user is not a derogatory user, the derogatory reason of the predicted user can be used to improve some existing shortcomings or experiences in subsequent product development.

[0069] In the embodiments of the present application, the electronic device determines whether the predicted user is a derogatory user and determines the derogatory reason of the predicted user by obtaining the initial feedback samples of the surveyed users and the initial feedback samples corresponding to the surveyed users. Thus, the shortcomings existing in the product or business can be determined in time, so that the users can be retained by timely improvement.

[0070] In the embodiments of the present application, the electronic device obtains the initial feedback samples corresponding to the plurality of surveyed users, and extracts initial derogatory keywords from the initial feedback samples to obtain an initial derogatory keyword set; determines similar users similar to the predicted user from the plurality of surveyed users, and determines first feedback samples corresponding to the similar users in the initial feedback samples; determines first derogatory keywords appearing in the initial derogatory keyword set from the first feedback samples; determines the confidence of each first derogatory keyword in the first feedback samples; and determines the target derogatory keyword corresponding to the predicted user from the first derogatory keywords according to the confidence. Thus, the derogatory reason of the predicted user is determined according to the initial feedback samples of the surveyed users and the initial feedback samples corresponding to the surveyed users.

[0071] For a more detailed understanding of the data processing method provided in the embodiments of the present application, please continue to refer to Figure 2 , Figure 2 is a second flowchart of the data processing method provided in the embodiments of the present application. The data processing method can include the following steps:

[0072] 201. Obtain initial feedback samples corresponding to a plurality of surveyed users, and perform word segmentation extraction and / or filter stop words on the initial feedback samples to obtain first processing samples.

[0073] In some implementations, initial feedback samples from different survey users can be obtained through questionnaire surveys, either online or offline. For example, some evaluations from survey users can be used as initial feedback samples, or different options in the questionnaire can also be used as initial feedback samples.

[0074] Specifically, electronic devices can integrate questionnaire content from the same survey user within the same time period, such as extracting textual information to obtain an initial feedback sample. This initial feedback sample contains information from the survey user's evaluation of the product or service.

[0075] It should be noted that a single survey user can correspond to multiple initial feedback samples from different times. For example, a survey user might have completed a questionnaire in January and then again in May. The January questionnaire would generate one initial feedback sample for that survey user, and the May questionnaire would generate another initial feedback sample for that survey user.

[0076] Specifically, the electronic device can perform word segmentation and / or filter stop words on the initial feedback sample to obtain the first processed sample. Then, the first processed sample is input into the text processing model to output the initial derogatory keywords and obtain the initial derogatory keyword set.

[0077] For example, an electronic device can first acquire all the text in the initial feedback sample, and then perform word segmentation and extraction on all the text, such as using a word segmentation tool. The extracted keywords can be various words and short phrases. This results in the first processed sample.

[0078] For example, electronic devices can also set corresponding stop words, such as positive words like "good reviews," "not bad," and "easy to use," which have positive connotations. These words can be set as stop words, and then all text can be filtered out by these stop words, thus obtaining the first processed sample.

[0079] In the process of processing the initial feedback sample, the electronic device can first extract the text of the initial feedback sample to obtain the corresponding multiple text of the initial feedback sample, then filter out some words by filtering out stop words, and then extract the text after filtering out stop words by word segmentation to obtain the first processed sample.

[0080] 202. Input the first processed sample into the text processing model, output the initial derogatory keywords, and obtain the initial derogatory keyword set.

[0081] The first processing sample is processed by a text processing model to obtain initial derogatory keywords, so as to form an initial derogatory keyword set according to the initial derogatory keywords. For example, the text processing model can obtain the word vector of different keywords in the first processing sample, then generate the initial derogatory keywords according to the word vector, and finally generate the initial derogatory keyword set.

[0082] The text processing model can use a Bert model or a keyBert model to complete keyword extraction of the first processing sample and obtain initial derogatory keywords. The initial derogatory keywords are words that the research user generates negative evaluations on products or services.

[0083] 203、Obtain the first feature corresponding to each research user and the second feature of the predicted user.

[0084] Each user has corresponding user features, and the user features can be user age, gender, consumption habits, economic strength, personal preferences, and other multi-dimensional user features.

[0085] For example, the electronic device can determine the user features of each research user according to the personal data of each research user in the database, such as obtaining the personal data authorized by the user for storage and use in the database, such as age, gender, consumption habits, economic strength, personal preferences, and other multi-dimensional user features. Then generate a corresponding vector for each dimension of the research user feature, and finally generate the first feature corresponding to the research user according to the vector of each dimension of the user feature. The first feature can be understood as the vector addition of the multi-dimensional research user feature.

[0086] The electronic device can determine the user features of the predicted user according to the personal data of each predicted user in the database, such as obtaining the personal data authorized by the user for storage and use in the database, such as age, gender, consumption habits, economic strength, personal preferences, and other multi-dimensional user features. Then generate a corresponding vector for each dimension of the predicted user feature, and finally generate the second feature corresponding to the predicted user according to the vector of each dimension of the user feature. The second feature can be understood as the vector addition of the multi-dimensional user feature.

[0087] 204、According to the first feature and the second feature, determine the similarity between each research user and the predicted user, and determine the research user with a similarity greater than a preset similarity threshold as a similar user.

[0088] Finally, the electronic device can obtain the feature similarity between each first feature and second feature, such as using cosine distance, Euclidean distance, Pearson correlation coefficient, and other ways to obtain the feature similarity between each first feature and second feature.

[0089] The electronic device can determine the feature similarity between each first feature and second feature as the similarity between each investigation user and the prediction user, and finally determine the investigation user with a similarity greater than a preset similarity threshold as a similar user.

[0090] In some embodiments, the electronic device can further input the first feature and second feature of each investigation user into a similarity calculation formula to obtain the similarity corresponding to each investigation user, and the similarity calculation formula is:

[0091]

[0092] wherein s j is the similarity, a is the attenuation coefficient, j is the investigation user, x is the prediction user, diff(j) is the time difference between the current date and the investigation date of the investigation user, is the cosine similarity between the investigation user and the prediction user.

[0093] wherein one investigation user can be investigated multiple times, such as once in January and once in May, and the investigation sample corresponding to the investigation time closest to the current time can be taken as the feedback sample of the investigation user.

[0094] In the above similarity calculation formula, the smaller the time difference between the current date and the investigation date of the investigation user, the greater the value of the first half of the multiplication in the similarity calculation formula, i.e. The greater the value of The greater the similarity value corresponding to the investigation user.

[0095] 205、Determine the first feedback sample corresponding to the similar user in the initial feedback sample, and calculate the net recommendation value corresponding to each similar user according to the first feedback sample corresponding to each similar user.

[0096] In some embodiments, the electronic device can calculate the net recommendation value corresponding to each similar user by using the corresponding NPS algorithm and the first feedback sample corresponding to each similar user.

[0097] For example, the net recommendation value can be a value in the range of 0-10.

[0098] 206、Determine the first derogatory user among the similar users according to the net recommendation value.

[0099] In some embodiments, the user with a net recommendation value less than a preset net recommendation value is determined as the first derogatory user, such as the preset net recommendation value can be set to 6, as long as the net recommendation value of the similar user is less than 6, the user is the first derogatory user.

[0100] 207、determine the proportion of the first derogatory user in the similar users, and if the proportion is greater than a preset proportion threshold, determine that the prediction user is a target derogatory user.

[0101] That is, when the proportion of the first derogatory user in the similar users is greater than the preset proportion threshold, it indicates that most of the similar users of the prediction user are derogatory users, and the prediction user is also a derogatory user with a high probability.

[0102] 208、determine the first derogatory keyword in the initial derogatory keyword set in the first feedback sample.

[0103] In some embodiments, the electronic device can further obtain each keyword of the first feedback sample after determining the first feedback sample corresponding to the similar user, and then determine the keyword appearing in the initial derogatory keyword set as the first derogatory keyword.

[0104] For example, the electronic device can match each keyword with the initial derogatory keyword set by keyword matching, and if a keyword matches successfully, the keyword is determined as the first derogatory keyword.

[0105] The electronic device can further determine the number of times each first derogatory keyword appears in the first feedback sample after determining the first derogatory keyword. For example, a first derogatory keyword can be matched with each first feedback sample, and the number of times the first derogatory keyword appears in each first feedback sample is counted, and finally the numbers are added to obtain the number of times the first derogatory keyword appears in all first feedback samples.

[0106] 209、determine the confidence of each first derogatory keyword in the first feedback sample.

[0107] In some embodiments, the electronic device can determine a target first derogatory keyword in the first derogatory keyword, determine a first total number of times the target first derogatory keyword appears in all first feedback samples, determine a second total number of times all first derogatory keywords appear in all first feedback samples, and finally divide the first total number by the second total number to obtain the confidence corresponding to the target first derogatory keyword.

[0108] For example, the target first derogatory keyword includes keyword A, keyword B, and keyword C, wherein keyword A is the target first derogatory keyword, and the first total number of times keyword A appears in all first feedback samples is obtained. The electronic device further determines the second total number of times keyword A, keyword B, and keyword C appear in all first feedback samples. Finally, the first total number is divided by the second total number to obtain the confidence corresponding to keyword A.

[0109] In some embodiments, the electronic device can further determine first similar users corresponding to the target first derogatory keyword, and then obtain a similarity between each first similar user and the predicted user, add the similarity corresponding to each first similar user, obtain a first sum of similarities, and multiply the first sum of similarities by a first total number of occurrences of the target first derogatory keyword in all first feedback samples to obtain a first product result.

[0110] The electronic device obtains a second sum of similarities of similar users corresponding to all first derogatory keywords, and a second sum of similarities corresponding to each similar user and the predicted user, and then multiplies the second sum of similarities by a second total number of occurrences of all first derogatory keywords in all first feedback samples to obtain a second product result.

[0111] Finally, the first product result is divided by the second product result to obtain the confidence of the target first derogatory keyword.

[0112] In this way, the confidence corresponding to each first keyword in all first feedback samples can be determined in sequence.

[0113] 210、In the first derogatory keywords, a second derogatory keyword with a confidence greater than a preset confidence is determined, and the second derogatory keyword is determined as the target derogatory keyword corresponding to the predicted user.

[0114] For example, the electronic device can preset a preset confidence threshold. When the confidence of a first derogatory keyword is greater than the preset confidence threshold, it means that the first derogatory keyword has a higher confidence, and the first derogatory keyword is confirmed as the second derogatory keyword, i.e., the target derogatory keyword corresponding to the predicted user.

[0115] In the embodiments of the present application, the electronic device determines whether the predicted user is a derogatory user and determines the derogatory reason corresponding to the predicted user by obtaining the initial feedback sample of the investigated user and the initial feedback sample corresponding to the investigated user. Thus, the shortcomings existing in the product or service can be determined in time, so that the user can be retained by timely improvement.

[0116] In the embodiments of the present application, the electronic device obtains the initial feedback sample corresponding to the plurality of investigated users, performs word segmentation extraction and / or filtering of stop words on the initial feedback sample to obtain a first processed sample. The first processed sample is input into a text processing model to output initial derogatory keywords to obtain an initial derogatory keyword set.

[0117] Then, a first feature corresponding to each investigation user and a second feature of the predicted user are obtained. Similarity between each investigation user and the predicted user is determined according to the first feature and the second feature, and an investigation user with similarity greater than a preset similarity threshold is determined as a similar user. A first feedback sample corresponding to the similar user in the initial feedback sample is determined, and a net recommendation value corresponding to each similar user is calculated according to the first feedback sample corresponding to each similar user. A first derogatory user is determined among the similar users according to the net recommendation value.

[0118] Finally, a proportion of the first derogatory user in the similar users is determined, and if the proportion is greater than a preset proportion threshold, the predicted user is determined as a target derogatory user. A first derogatory keyword appearing in the initial derogatory keyword set is determined in the first feedback sample. A confidence degree of each first derogatory keyword in the first feedback sample is determined. A second derogatory keyword with a confidence degree greater than a preset confidence degree is determined among the first derogatory keywords, and the second derogatory keyword is determined as a target derogatory keyword corresponding to the predicted user. Thus, the derogatory reason of the predicted user is determined according to the investigation users and the initial feedback samples corresponding to the investigation users.

[0119] Please refer to Figure 3 , Figure 3 is a structural schematic diagram of a data processing apparatus provided by the embodiment of the present application. The data processing apparatus 300 can include:

[0120] The obtaining module 310 is configured to obtain initial feedback samples corresponding to a plurality of investigation users, and perform initial derogatory keyword extraction on the initial feedback samples to obtain an initial derogatory keyword set.

[0121] The obtaining module 310 is further configured to perform word segmentation extraction and / or filter stop words on the initial feedback samples to obtain a first processing sample; and input the first processing sample into a text processing model to output the initial derogatory keyword and obtain the initial derogatory keyword set.

[0122] The first determining module 320 is configured to determine similar users similar to the predicted user among the plurality of investigation users, and determine first feedback samples corresponding to the similar users in the initial feedback samples.

[0123] The first determining module 320 is further configured to calculate a net recommendation value corresponding to each similar user according to the first feedback sample corresponding to each similar user.

[0124] The first derogatory user is determined among the similar users according to the net recommendation value.

[0125] The first determining module 320 is further configured to determine a target similar user with a net recommendation value less than a preset net recommendation value among the similar users.

[0126] The target similar user is determined as the first derogatory user.

[0127] The first determining module 320 is further configured to determine a proportion of the first derogatory user in similar users.

[0128] If the proportion is greater than a preset proportion threshold, the prediction user is determined as a target derogatory user.

[0129] The first determining module 320 is further configured to obtain a first feature corresponding to each investigation user and a second feature of the prediction user.

[0130] The first determining module 320 is further configured to determine a similarity between each investigation user and the prediction user according to the first feature and the second feature.

[0131] The first determining module 320 is further configured to determine an investigation user with a similarity greater than a preset similarity threshold as a similar user.

[0132] The first determining module 320 is further configured to input the first feature and the second feature of each investigation user into a similarity calculation formula to obtain a similarity corresponding to each investigation user, and the similarity calculation formula is as follows:

[0133]

[0134] wherein, s j is the similarity, a is a decay coefficient, j is the investigation user, x is the prediction user, diff(j) is a time difference between a current date and a survey date of the investigation user, is a cosine similarity between the investigation user and the prediction user.

[0135] The second determining module 330 is configured to determine a first derogatory keyword appearing in the initial derogatory keyword set in the first feedback sample.

[0136] The third determining module 340 is configured to determine a confidence degree of each first derogatory keyword in the first feedback sample.

[0137] The third determining module 340 is further configured to determine a target first derogatory keyword in the first derogatory keywords, and determine a first total number of times that the target first derogatory keyword appears in all the first feedback samples.

[0138] The third determining module 340 is further configured to determine a second total number of times that all the first derogatory keywords appear in all the first feedback samples.

[0139] The third determining module 340 is further configured to divide the first total number by the second total number to obtain a confidence degree corresponding to the target first derogatory keyword.

[0140] The fourth determining module 350 is configured to determine a target derogatory keyword corresponding to the prediction user in the first derogatory keywords according to the confidence degree.

[0141] The fourth determination module 350 is further configured to determine, from the first derogatory keywords, a second derogatory keyword with a confidence degree greater than a preset confidence degree, and determine the second derogatory keyword as the target derogatory keyword corresponding to the predicted user.

[0142] In the embodiments of the present application, the electronic device obtains initial feedback samples corresponding to a plurality of investigation users, and extracts initial derogatory keywords from the initial feedback samples to obtain an initial derogatory keyword set. Similar users similar to the predicted user are determined from the plurality of investigation users, and a first feedback sample corresponding to the similar user in the initial feedback samples is determined. First derogatory keywords appearing in the initial derogatory keyword set are determined from the first feedback sample. The confidence degree of each first derogatory keyword in the first feedback sample is determined. The target derogatory keyword corresponding to the predicted user is determined from the first derogatory keywords according to the confidence degree. Thus, the derogatory reason of the predicted user is determined according to the investigation users and the initial feedback samples corresponding to the investigation users.

[0143] Correspondingly, the embodiments of the present application also provide an electronic device, as shown in the Figure 4 The electronic device 400 can include a memory 401 having one or more computer readable storage media, an input unit 402, a display unit 403, a sensor 404, a processor 405 including one or more processing cores, and a power supply 406, and the like. Those skilled in the art can understand that the structure of the electronic device shown in Figure 4 The structure of the electronic device shown in the

[0144] The memory 401 can be used to store software programs and modules, and the processor 405 can execute various functions and data processing by running the software programs and modules stored in the memory 401. The memory 401 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), and the like; the data storage area can store data created according to the use of the electronic device (such as audio data, a phone book, etc.), and the like. In addition, the memory 401 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device. Accordingly, the memory 401 can also include a memory controller to provide access of the processor 405 and the input unit 402 to the memory 401.

[0145] The input unit 402 can be configured to receive input digital or character information, and generate a key signal, a mouse signal, a joystick signal, an optical or a trackball signal associated with a user's setting and function control. Specifically, in one embodiment, the input unit 402 can include a touch-sensitive surface and other input devices. The touch-sensitive surface, also known as a touch display or a touchpad, can collect a user's touch operation (e.g., the user's operation on or near the touch-sensitive surface using a finger, a stylus, or any suitable object or accessory) and drive a corresponding connection device according to a pre-set program. Optionally, the touch-sensitive surface can include two parts, a touch detection device and a touch controller. The touch detection device detects the user's touch position and detects a signal caused by the touch operation, and transmits the signal to the touch controller. The touch controller receives the touch information from the touch detection device, converts it into touch coordinates, and sends it to the processor 405, and can also receive commands from the processor 405 and execute them. In addition, the touch-sensitive surface can be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave. In addition to the touch-sensitive surface, the input unit 402 can also include other input devices. Specifically, the other input devices can include one or more of a physical keyboard, function keys (such as volume control keys, on / off keys, etc.), a trackball, a mouse, a joystick, etc.

[0146] The display unit 403 can be configured to display information input by a user or information provided to the user, and various graphical user interfaces of the electronic device, which can be composed of graphics, text, icons, video, and any combination thereof. The display unit 403 can include a display panel, which can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. Further, the touch-sensitive surface can cover the display panel, and when the touch-sensitive surface detects a touch operation on or near it, it transmits to the processor 405 to determine the type of touch event, and then the processor 405 provides corresponding visual output on the display panel according to the type of touch event. Although in the above embodiment, the touch-sensitive surface and the display panel are implemented as two independent components to realize input and output functions, in some embodiments, the touch-sensitive surface and the display panel can be integrated to realize input and output functions. Figure 4

[0147] ​The electronic device can further include at least one sensor 404, such as a light sensor, a motion sensor, and other sensors. Specifically, the light sensor can include an ambient light sensor and a proximity sensor, where the ambient light sensor can adjust the brightness of the display panel according to the brightness of ambient light, and the proximity sensor can turn off the display panel and / or backlight when the electronic device is moved to the ear. As one of the motion sensors, the gravity acceleration sensor can detect the magnitude of acceleration in each direction (generally three axes), and when at rest, it can detect the magnitude and direction of gravity, and can be used for applications that identify the posture of the electronic device (such as landscape / portrait switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer, tapping), and the like. As for other sensors that the electronic device can also be configured, such as a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, and the like, they will not be described here.

[0148] The processor 405 is the control center of the electronic device, which connects various parts of the entire electronic device through various interfaces and lines, and performs various functions of the electronic device and processes data by running or executing software programs and / or modules stored in the memory 401 and calling data stored in the memory 401, thereby overall monitoring the electronic device. Optionally, the processor 405 can include one or more processing cores; preferably, the processor 405 can integrate an application processor and a modem processor, where the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 405.

[0149] The electronic device further includes a power supply 406 (such as a battery) for supplying power to various components, and preferably, the power supply can be logically connected to the processor 405 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 406 can also include one or more direct or alternating current power supplies, a recharging system, a power failure detection circuit, a power converter or inverter, a power state indicator, and any other components.

[0150] Although not shown, the electronic device can also include a camera, a Bluetooth module, and the like, which will not be described here. Specifically, in the present embodiment, the processor 405 in the electronic device loads the computer program stored on the memory 401, and the processor 405 realizes various functions by loading the computer program:

[0151] Obtaining a plurality of initial feedback samples corresponding to investigation users, and performing initial derogatory keyword extraction on the initial feedback samples to obtain an initial derogatory keyword set;

[0152] determine similar users similar to the predicted user from the plurality of investigation users, and determine first feedback samples corresponding to the similar users in the initial feedback samples;

[0153] determine first derogatory keywords appearing in the initial derogatory keyword set in the first feedback samples;

[0154] determine a confidence degree of each first derogatory keyword in the first feedback samples;

[0155] determine a target derogatory keyword corresponding to the predicted user from the first derogatory keywords according to the confidence degrees.

[0156] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions or by controlling related hardware by the instructions, which can be stored in a computer readable storage medium and loaded and executed by a processor.

[0157] To this end, the embodiments of the present application provide a computer readable storage medium, which stores a plurality of instructions capable of being loaded by a processor to execute steps in any data processing method provided by the embodiments of the present application. For example, the instructions can execute the following steps:

[0158] obtain initial feedback samples corresponding to a plurality of investigation users, and perform initial derogatory keyword extraction on the initial feedback samples to obtain an initial derogatory keyword set;

[0159] determine similar users similar to the predicted user from the plurality of investigation users, and determine first feedback samples corresponding to the similar users in the initial feedback samples;

[0160] determine first derogatory keywords appearing in the initial derogatory keyword set in the first feedback samples;

[0161] determine a confidence degree of each first derogatory keyword in the first feedback samples;

[0162] determine a target derogatory keyword corresponding to the predicted user from the first derogatory keywords according to the confidence degrees.

[0163] The specific implementation of each operation can refer to the foregoing embodiments, which will not be described here.

[0164] The storage medium can include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0165] Since the instructions stored in the storage medium can execute the steps in any one of the data processing methods provided by the embodiments of the present application, the beneficial effects that can be achieved by any one of the data processing methods provided by the embodiments of the present application can be achieved. Details are described above, and will not be repeated here.

[0166] The above describes in detail a data processing method, device, electronic equipment and storage medium provided by the embodiments of the present application. The principles and implementation manners of the present application are described by applying specific examples. The above description of the embodiments is only used to help understand the method and its core idea of the present application. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A data processing method, characterized by, The method comprises the following steps: obtaining initial feedback samples corresponding to a plurality of investigation users, and performing initial derogatory keyword extraction on the initial feedback samples to obtain an initial derogatory keyword set; the initial derogatory keyword is a word that the investigation user produces a negative evaluation on a product or service; determining similar users similar to a prediction user from the plurality of investigation users, and determining first feedback samples corresponding to the similar users in the initial feedback samples; determining first derogatory keywords appearing in the initial derogatory keyword set from the first feedback samples; determining a confidence degree of each first derogatory keyword in the first feedback samples; determining a target derogatory keyword corresponding to the prediction user from the first derogatory keywords according to the confidence degree; calculating a net recommendation value corresponding to each similar user according to the first feedback sample corresponding to each similar user; determining a first derogatory user from the similar users according to the net recommendation value; determining a proportion of the first derogatory user in the similar users; if the proportion is greater than a preset proportion threshold, determining that the prediction user is a target derogatory user.

2. The data processing method according to claim 1, characterized in that, The method of determining a first derogatory user from the similar users according to the net recommendation value comprises: determining a target similar user with a net recommendation value less than a preset net recommendation value from the similar users; determining the target similar user as the first derogatory user.

3. The data processing method of claim 1, wherein, The method of performing initial derogatory keyword extraction on the initial feedback samples to obtain an initial derogatory keyword set comprises: performing word segmentation extraction and / or filtering stop words on the initial feedback samples to obtain a first processing sample; inputting the first processing sample into a text processing model to output the initial derogatory keywords and obtain the initial derogatory keyword set.

4. The data processing method of claim 1, wherein, The method of determining similar users similar to a prediction user from the plurality of investigation users comprises: obtaining a first feature corresponding to each investigation user and a second feature of the prediction user; determining a similarity between each investigation user and the prediction user according to the first feature and the second feature; determining the investigation user with a similarity greater than a preset similarity threshold as the similar user.

5. The data processing method according to claim 4, characterized in that, The method of determining a similarity between each investigation user and the prediction user according to the first feature and the second feature comprises: inputting the first feature and the second feature of each investigation user into a similarity calculation formula to obtain a similarity corresponding to each investigation user, and the similarity calculation formula is: wherein s j is the similarity, a is the decay coefficient, j is the survey user, x is the predicted user, diff(j) is the time difference between the current date and the survey date of the survey user, is the cosine similarity between the survey user and the predicted user.

6. The data processing method of claim 1, wherein, The method of determining a confidence degree of each first derogatory keyword in the first feedback samples comprises: determining a target first derogatory keyword from the first derogatory keywords, and determining a first total number of times that the target first derogatory keyword appears in all the first feedback samples; determining a second total number of times that all the first derogatory keywords appear in all the first feedback samples; dividing the first total number of times by the second total number of times to obtain a confidence degree corresponding to the target first derogatory keyword.

7. The data processing method according to any one of claims 1 to 6, characterized in that, The target derogatory keyword corresponding to the predicted user is determined from the first derogatory keyword according to the confidence. A second derogatory keyword with a confidence greater than a preset confidence is determined from the first derogatory keyword, and the second derogatory keyword is determined as the target derogatory keyword corresponding to the predicted user.

8. A data processing apparatus, characterized by, Comprise: An acquisition module is configured to acquire initial feedback samples corresponding to a plurality of investigation users, and perform initial derogatory keyword extraction on the initial feedback samples to obtain an initial derogatory keyword set; The initial derogatory keyword is a word used by an investigation user to make a negative evaluation on a product or a service; A first determination module is configured to determine similar users similar to a predicted user from the plurality of investigation users, and determine first feedback samples corresponding to the similar users in the initial feedback samples; A second determination module is configured to determine first derogatory keywords appearing in the initial derogatory keyword set from the first feedback samples; A third determination module is configured to determine a confidence of each of the first derogatory keywords in the first feedback samples; A fourth determination module is configured to determine a target derogatory keyword corresponding to the predicted user from the first derogatory keywords according to the confidence; The first determination module is further configured to perform the following operations: Calculate a net recommendation value corresponding to each of the similar users according to the first feedback sample corresponding to each of the similar users; Determine a first derogatory user from the similar users according to the net recommendation value; Determine a proportion of the first derogatory user in the similar users; If the proportion is greater than a preset proportion threshold, determine the predicted user as a target derogatory user.

9. An electronic device, comprising: Comprise: A memory storing executable program codes, and a processor coupled with the memory; The processor invokes the executable program codes stored in the memory to execute steps in the data processing method according to any one of claims 1 to 7.

10. A storage medium, characterized by The storage medium stores a plurality of instructions, and the instructions are adapted to be loaded by the processor to execute steps in the data processing method according to any one of claims 1 to 7.

Images