CN113988948A

CN113988948A - Satisfaction collecting method, device, equipment and storage medium

Info

Publication number: CN113988948A
Application number: CN202111340198.4A
Authority: CN
Inventors: 马国斌; 杨晓; 梁寿亮; 庞文强; 王丽云
Original assignee: Bank of China Ltd
Current assignee: Bank of China Ltd
Priority date: 2021-11-12
Filing date: 2021-11-12
Publication date: 2022-01-28

Abstract

The application provides a satisfaction collecting method, a satisfaction collecting device, satisfaction collecting equipment and a storage medium, and relates to the technical field of big data. The method comprises the following steps: acquiring first log data of a plurality of users; wherein the first log data includes: the method comprises the steps that an access record and a user identification are obtained, wherein the access record comprises a plurality of page URLs with time sequence generated when a user accesses a target application within a preset time period by using a terminal; for each user of a plurality of users: determining transaction information of the user according to the first log data of the user and the service scene classification model, and acquiring second log data of the user according to the identification of the user and the transaction information of the user; and determining the user satisfaction degree of the target service according to the transaction information and the second log data of each user in the plurality of users. The method and the device can be used in the process of collecting the user satisfaction.

Description

Satisfaction collecting method, device, equipment and storage medium

Technical Field

The present application relates to the field of big data processing, and in particular, to a satisfaction collecting method, apparatus, device, and storage medium.

Background

In the modern society of internet financial impact, the banking competition is increasingly intensified. For banks, it is important to collect the satisfaction of users on their products. And the user satisfaction is acquired in time, and the service strategy is adjusted according to users with different satisfaction degrees, so that the bank can maintain the users, and the loss of the users is avoided.

In the prior art, the collection method of user satisfaction is generally divided into two methods. One is to measure the potential satisfaction of the user with the product by means of a questionnaire. And secondly, collecting evaluation information of the user through a big data mining technology, and acquiring the satisfaction degree of the user by adopting an emotion analysis method. The two satisfaction degree collection methods are both based on subjective opinions of users, and due to the fact that the difference of client cognition and the variability of the subjective attitudes of the users exist in collected data, the obtained user satisfaction degree is not accurate and objective enough, so that a bank cannot effectively adjust a service strategy according to the user satisfaction degree, poor use experience is easily brought to the users, and the users are lost.

Disclosure of Invention

The application provides a satisfaction collecting method, device, equipment and storage medium, which can accurately and objectively collect user satisfaction so as to adjust a service strategy according to the user satisfaction and bring better use experience to the user.

In a first aspect, the present application provides a satisfaction collection method, comprising: acquiring first log data of a plurality of users; wherein the first log data includes: the method comprises the steps that an access record and a user identification are obtained, wherein the access record comprises a plurality of page Uniform Resource Locators (URLs) with time sequence, which are generated when a user accesses a target application within a preset time period by using a terminal; for each user of a plurality of users: determining transaction information of the user according to the first log data of the user and the service scene classification model, and acquiring second log data of the user according to the identification of the user and the transaction information of the user; wherein, the transaction information of the user comprises: one or more services transacted by a user through a target application in a preset time period and an access time period corresponding to each service; the second log data includes: information whether each service transacted by the user through the target application is successful or not in an access time period corresponding to each service; determining user satisfaction of the target service according to the transaction information and the second log data of each user in the plurality of users; the target service comprises all services transacted by a plurality of users through the target application within a preset time period.

In one possible implementation manner, determining the user satisfaction of the target service according to the transaction information and the second log data of each of the plurality of users includes: aiming at each user in the plurality of users, determining the transaction success rate of each service transacted by the user through the target application within a preset time period according to the second log data of the user; determining the user satisfaction degree of the target service according to the service transacted by the target application in the transaction information of the plurality of users within a preset time period and the transaction success rate of each service transacted by each user within the preset time period; wherein, the transaction success rate of the business transaction is in direct proportion to the user satisfaction.

In another possible implementation manner, the method further includes: acquiring a training data set; the training data set comprises a plurality of pieces of training data, one piece of training data comprises an access record and corresponding service information, wherein the access record is generated when a sample user uses a terminal to access a target application to transact a service, and the access record comprises a plurality of page URLs with time sequence; the service information corresponding to the access record is self-defined; performing model training according to the training data set to obtain a business scene classification model; the service scene classification model has the function of determining the transacted service according to the access record.

In another possible implementation manner, in the case that the business transaction transacted by the user through the target application fails, the second log data further includes: a transaction failure reason; the method further comprises the following steps: and determining the dissatisfaction reason of the target service user according to the transaction information and the second log data of each user in the plurality of users.

In another possible implementation manner, performing model training according to a training data set to obtain a business scenario classification model includes: and inputting the training data set into a Recurrent Neural Network (RNN) model for training to obtain a business scene classification model.

According to the satisfaction collecting method provided by the application, the satisfaction collecting device can determine the business transacted by the user and the corresponding access time period of transacting the business according to the access record in the foreground log data of the user. Then, the satisfaction collecting device can acquire the information whether the transaction of the service transacted by the user is successful from the background log data according to the access time interval and the identification of the user, and determine the user satisfaction of each service according to the service transacted by the user and the information whether the transaction of the transacted service is successful. The satisfaction collecting method is used for determining the user satisfaction of each service based on the service transacted by all users and the information whether all users in each service transact successfully or not, which are obtained according to log data, has objective accuracy, is not influenced by the subjective attitude of the users, and ensures that the collected user satisfaction is more reliable, so that a bank can adjust a service strategy according to the user satisfaction, better use experience is brought to the users, and loss of the users is avoided.

Furthermore, the method adopts the RNN model to determine the business which the user wants to handle, the RNN model has good effect of processing the sequence data, the obtained data is more accurate, and the collected user satisfaction degree is more reliable. In addition, the method can also determine the dissatisfaction reasons of users of different services, so as to adjust the service strategy in time according to the dissatisfaction reasons of the users and better maintain the users.

In a second aspect, the present application provides a satisfaction collecting device, comprising: the device comprises an acquisition module and a determination module; the acquisition module is used for acquiring first log data of a plurality of users; wherein the first log data includes: the method comprises the steps that an access record and a user identification are obtained, wherein the access record comprises a plurality of page URLs with time sequence generated when a user accesses a target application within a preset time period by using a terminal; for each user of a plurality of users: the acquisition module is further used for determining the transaction information of the user according to the first log data of the user and the service scene classification model, and acquiring second log data of the user according to the identification of the user and the transaction information of the user; wherein, the transaction information of the user comprises: one or more services transacted by a user through a target application in a preset time period and an access time period corresponding to each service; the second log data includes: information whether each service transacted by the user through the target application is successful or not in an access time period corresponding to each service; the determining module is used for determining the user satisfaction degree of the target service according to the transaction information and the second log data of each user in the plurality of users; the target service comprises all services transacted by a plurality of users through the target application within a preset time period.

In a possible implementation manner, the determining module is specifically configured to: aiming at each user in the plurality of users, determining the transaction success rate of each service transacted by the user through the target application within a preset time period according to the second log data of the user; determining the user satisfaction degree of the target service according to the service transacted by the target application in the transaction information of the plurality of users within a preset time period and the transaction success rate of each service transacted by each user within the preset time period; wherein, the transaction success rate of the business transaction is in direct proportion to the user satisfaction.

In another possible implementation manner, the apparatus further includes: establishing a module; an establishment module to: acquiring a training data set; acquiring a training data set; the training data set comprises a plurality of pieces of training data, one piece of training data comprises an access record and corresponding service information, wherein the access record is generated when a sample user uses a terminal to access a target application to transact a service, and the access record comprises a plurality of page URLs with time sequence; the service information corresponding to the access record is self-defined; performing model training according to the training data set to obtain a business scene classification model; the service scene classification model has the function of determining the transacted service according to the access record.

In yet another possible implementation manner, the determining module is further configured to determine a user dissatisfaction reason of the target service according to the transaction information and the second log data of each of the plurality of users.

In another possible implementation manner, the establishing module is specifically configured to input the training data set into the RNN model for training, so as to obtain a service scene classification model.

In a third aspect, the present application provides an electronic device comprising: a processor and a memory; the memory stores instructions executable by the processor; the processor is configured to execute the instructions such that the electronic device implements the method of the first aspect described above.

In a fourth aspect, the present application provides a computer-readable storage medium comprising: computer software instructions; the computer software instructions, when executed in an electronic device, cause the electronic device to carry out the method of the first aspect described above.

In a fifth aspect, the present application provides a computer program product for causing a computer to perform the steps of the related method described in the above first aspect, when the computer program product runs on a computer, so as to implement the method of the above first aspect.

The beneficial effects of the second to fifth aspects may refer to the corresponding description of the first aspect, and are not repeated.

Drawings

FIG. 1 is a schematic diagram of an application environment of a satisfaction degree collection method provided by the present application;

FIG. 2 is a schematic flow chart of a satisfaction degree collection method provided by the present application;

FIG. 3 is a schematic flow chart of another satisfaction collection method provided herein;

FIG. 4 is a schematic diagram of an RNN model training process provided herein;

FIG. 5 is a schematic diagram of a conventional neural network and RNN provided herein;

FIG. 6 is a schematic flow chart of another satisfaction collection method provided herein;

FIG. 7 is a schematic diagram of a satisfaction gathering device provided herein;

FIG. 8 is a schematic diagram of the composition of another satisfaction gathering device provided herein;

fig. 9 is a schematic composition diagram of an electronic device provided in the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

It should be noted that in the embodiments of the present application, words such as "exemplary" or "for example" are used to mean serving as examples, illustrations or descriptions. Any embodiment or design described herein as "exemplary" or "e.g.," is not necessarily to be construed as preferred or advantageous over other embodiments or designs. Rather, use of the word "exemplary" or "such as" is intended to present concepts related in a concrete fashion.

For the convenience of clearly describing the technical solutions of the embodiments of the present application, in the embodiments of the present application, the terms "first", "second", and the like are used for distinguishing the same items or similar items with basically the same functions and actions, and those skilled in the art can understand that the terms "first", "second", and the like are not limited in number or execution order.

At present, it is very important for banks to collect the satisfaction degree of users on their products. And the user satisfaction is acquired in time, and the service strategy is adjusted according to users with different satisfaction degrees, so that the bank can maintain the users, and the loss of the users is avoided.

The traditional method for collecting the user satisfaction degree comprises the steps of passively receiving customer complaints and actively communicating with customers directly, the collection period of the method is long, and the real-time performance and the comprehensiveness of the method for collecting the user satisfaction degree of the internet application are not achieved for hundred million user quantities. Therefore, in the prior art, the internet industry generally adopts a Net Promoter Score (NPS) method and a Net Word-Of-Mouth Rate (NWR) method to collect customer satisfaction.

The NPS method measures the potential satisfaction degree of a user on a product in a questionnaire mode, and due to the difference in cognition of the user, the collected data has distortion. The NWR method processes the evaluation information of a user through a big data mining technology, directly obtains word-of-mouth information related to a specified product within a certain period of time by means of word segmentation, emotion interpretation and the like, and then judges the satisfaction degree of the user. The public praise information obtained by the NWR method is directly issued by a client, and has larger subjectivity, so that the user satisfaction obtained by analysis is deviated from the real situation. Moreover, the obtained data volume is related to the application usage and the promotion strength of the promotion program, so that the problem that the obtained data is not comprehensive enough exists. Meanwhile, the method passively depends on the client to release the evaluation information, the time for obtaining the result is long, and the real-time performance is not achieved. Therefore, the NPS method and the NWR method have the problems that the result is not accurate and objective enough, the data is not comprehensive enough, and the real-time acquisition cannot be realized.

In this background technology, embodiments of the present application provide a satisfaction collecting method, by which more accurate and objective user satisfaction can be comprehensively collected in real time, so that a bank can adjust a service policy in a targeted manner according to the obtained user satisfaction, and loss of users is avoided.

The satisfaction degree collecting method provided by the application can be applied to the application environment shown in the figure 1. As shown in fig. 1, the application environment may include: a satisfaction degree collecting device 101, a front-end apparatus 102, and a back-end apparatus 103. The satisfaction collecting means 101, the front-end apparatus 102 and the back-end apparatus 103 are connected to each other.

The satisfaction collecting device 101 may be a server cluster composed of a plurality of servers, or a single server, or a computer, or a processor or a processing chip in a server or a computer, or the like. The embodiment of the present application does not limit the specific apparatus form of the satisfaction degree collecting device 101. The satisfaction collecting means 101 is illustrated as a single server in fig. 1.

The front-end device 102 may be an electronic device such as a mobile phone terminal, a computer, or an intelligent counter. The embodiment of the present application does not limit the specific device form of the front-end device 102, and fig. 1 illustrates the front-end device as a mobile phone terminal as an example. The backend device 103 may be a server cluster composed of a plurality of servers, or a single server (as shown in fig. 1). The embodiment of the present application also does not limit the specific device form of the backend device 103. The user may transact business using the application in the front-end device 102. Accordingly, the backend device 103 may be an application server of the application, or a storage device connected to the application server of the application, and is configured to store relevant data of the application server, such as log data. The application (e.g., the target application in the present application) may be a bank application, such as a mobile banking application. For example, the user may use a cell phone bank to transact money transfer, fund purchase, etc. Of course, other applications supporting business handling, such as shopping applications, travel applications, etc., may also be provided.

In some embodiments, when the user accesses the target application using the front-end device 102, a plurality of Uniform Resource Locators (URLs) with time sequence are generated in the front-end log of the front-end device 102. When the user transacts the service using the target application, if the service transaction is successful, for example, the transaction is successful, a transaction code is generated in the back-end log of the back-end device 103 to indicate that the transaction is successful. If the transaction fails, the back-end log will generate the reason for the failure of the transaction. When it is necessary to collect the satisfaction of the user, the satisfaction collecting means 101 may acquire the front-end log from the front-end device 102. The satisfaction collecting device 101 can determine the services to be transacted by the user according to the front end log. Moreover, the satisfaction collecting device 101 may obtain the backend log from the backend device 103, and determine whether the service transacted by the user is successful according to the backend log. And then, determining the satisfaction degrees of the users in different services according to the user satisfaction degrees, so as to adjust the service strategy based on the user satisfaction degrees.

Fig. 2 is a schematic flowchart of a satisfaction degree collection method according to an embodiment of the present application. As shown in fig. 2, the satisfaction collecting method provided by the present application may be implemented by the satisfaction collecting apparatus described above, and specifically may include the following steps:

s201, first log data of a plurality of users are acquired.

The log is an event record generated during the operation of the network device, system or service program, and each row of the log records the description of the relevant operations such as date, time, user and action.

Wherein the first log data may include: and the access record comprises a plurality of page URLs with time sequence generated when the user accesses the target application within a preset time period by using the terminal and the identification of the user. The predetermined time period may be preset and may be set according to the requirements of the actual application scenario. The plurality of users may include users transacting business using the target application for a predetermined period of time.

In some embodiments, in the case that a user transacts a service using a target application in a terminal, the terminal generates and stores corresponding log data in the terminal. Based on this, the satisfaction collecting means may obtain the log data of the users, that is, the first log data, from the terminals of the users who transact business using the target application within a predetermined period of time to obtain the user's identification and a plurality of page URLs with time sequence generated when the user accesses the target application within the predetermined period of time using the terminal.

For each of the plurality of users in S201: s202 is performed as follows.

S202, determining the transaction information of the user according to the first log data of the user and the service scene classification model, and acquiring second log data of the user according to the identification of the user and the transaction information of the user.

The transaction information of the user may include: one or more services transacted by the user through the target application within a predetermined time period and an access period corresponding to each service. The second log data may include: and in the access time period corresponding to each service, the information whether each service transacted by the user through the target application is successful or not.

In some embodiments, it can be understood that, when the user transacts services using the terminal, the access records, that is, the accessed page URLs have a certain regularity, so that the main intention of the user, such as which service the user wants to transact, can be determined from the access records of the user. The access record in the obtained first log data (or referred to as front-end log data) includes page URLs generated when the user accesses the target application within a predetermined time period by using the terminal, and the URLs also have a time sequence (or the access record includes the access time of the page URLs), so that the satisfaction collecting device can determine one or more services handled by the user within the predetermined time period and the access time period corresponding to each service according to the first log data and the service scene classification model after obtaining the first log data of the user, that is, obtain the transaction information of the user.

In addition, when a user transacts a service by using a target application, whether the service is successfully transacted or not, log data (or called backend log data) is generated at a backend device to record whether the service is successfully transacted or not. Therefore, after determining the transaction information of the user, the satisfaction collecting device further obtains information whether the user successfully transacts within the access time period corresponding to each service from the back-end device according to the identification of the user, the services transacted by the user and the access time period of each service, that is, obtains the second log data.

S203, determining the user satisfaction degree of the target service according to the transaction information and the second log data of each user in the plurality of users.

The target service comprises all services transacted by a plurality of users through the target application within a preset time period.

In some embodiments, after obtaining the transaction information of the user and the second log data, the satisfaction collecting device may determine the user satisfaction of each service in all services transacted by the plurality of users within a preset time period according to the services transacted by each user in the plurality of users and information about whether each service is successfully transacted.

The technical scheme provided by the embodiment has the beneficial effects that the satisfaction degree collecting device can determine the business transacted by the user and the access time period corresponding to the business transacted according to the access record in the foreground log data of the user. Then, the satisfaction collecting device can acquire the information whether the transaction of the service transacted by the user is successful from the background log data according to the access time interval and the identification of the user, and determine the user satisfaction of each service according to the service transacted by the user and the information whether the transaction of the transacted service is successful. The satisfaction collecting method is used for determining the user satisfaction of each service based on the service transacted by all users and the information whether all users in each service transact successfully or not, which are obtained according to log data, has objective accuracy, is not influenced by the subjective attitude of the users, and ensures that the collected user satisfaction is more reliable, so that a bank can adjust a service strategy according to the user satisfaction, better use experience is brought to the users, and loss of the users is avoided.

A satisfaction collecting method provided by the embodiments of the present application will be described in detail below with reference to specific embodiments, and the satisfaction collecting method can be applied to a satisfaction collecting apparatus. The satisfaction collecting method may include two processes, i.e., a "model training process" and a "satisfaction collecting process", respectively. The "model training procedure" may be completed before the "satisfaction collection procedure".

As shown in FIG. 3, the "model training procedure" may include S301-S302 as follows. The "satisfaction collection procedure" may include S303 to S306 as follows.

S301, acquiring a training data set.

The training data set is used for indicating access records of a plurality of sample users and business information corresponding to each access record. The access record of the sample user may include a plurality of page URLs with a time sequence generated when the sample user transacts a service using the terminal access target application. The service information corresponding to the log data may be customized.

The satisfaction degree collecting device can obtain the visit records of a plurality of sample users and the business information corresponding to each visit record so as to obtain the training data set. The training data set may include a plurality of pieces of training data, and one piece of training data may include an access record of a sample user handling a service and corresponding service information. There may be a plurality of access records for one sample user. Each piece of training data may also include an identification of the user.

For example, the satisfaction collecting means may acquire, for each of a plurality of sample users, log data of the sample user from a terminal of the sample user. After the log data of the sample user is obtained, corresponding labels can be manually marked on access records generated by handling different services in the log data, for example, service information corresponding to the access records is marked. Then, the satisfaction collecting device can obtain the service information corresponding to the access record when the sample user transacts one or more services, so that one or more pieces of training data can be obtained, namely the training data is the access record with the service information. The pieces of training data of the plurality of sample users constitute a training data set. The service information may include a service scenario identifier, where the service scenario identifier is used to indicate a service that the user mainly wants to handle, and the service may be a transfer transaction, a purchase fund, or the like, for example, taking a target application as a bank application.

For example, if a user transacts a money transfer transaction needs to sequentially access pages A, B, and C, and a user purchasing a fund transaction needs to sequentially access pages A, D, and E. The satisfaction degree collecting means may acquire log data of the sample user. If the sample user 1 transacts the transfer transaction service, the access record in the log data of the terminal of the sample user 1 includes the URL of the page a-page B-page C with the time sequence, and then the corresponding service scene identifier can be manually marked on the access record: and (4) transferring the transaction. For another example, the sample user 2 transacts the basic service purchase, and the access record in the log data of the terminal of the sample user 2 includes the URL of the page a-page D-page E with the time sequence, then the corresponding service scene identifier may be manually marked for the access record: and (6) purchasing funds. After the manual marking is completed, the satisfaction collecting means can obtain training data 1 and training data 2. Wherein, the training data 1 is: sample user 1, page a-page B-page C, transfer transaction. Training data 2 was: sample user 2, page A-page D-page E, purchases funds. The training data 1 and the training data 2 constitute a training data set. It should be noted that, this example is illustrated by taking the training data set including the training data 1 and the training data 2 as an example, in practical application, the number of pieces of training data included in the training data set is not particularly limited, and the more the pieces of training data are, the better the pieces of training data are, so as to ensure the accuracy of the subsequent training model.

S302, inputting the training data set into a Recurrent Neural Network (RNN) model for training to obtain a business scene classification model.

The service scene classification model has the function of determining the transacted service according to the access record.

The satisfaction degree collecting device can train the model according to the training data set to obtain the business scene classification model.

In some embodiments, the satisfaction-gathering device may input the training data set into the RNN model for training to obtain the traffic scenario classification model.

For example, in connection with the above example, as shown in fig. 4, the satisfaction collection means inputs training data 1 and training data 2 into the RNN model, where the user identification and the access record in each training data are used as inputs, and the service scenario identification is used as an output for training. The trained RNN model is a business scene classification model, and the satisfaction degree collection device inputs the access records needing business scene classification into the business scene classification model so as to obtain the output result of the business scene identification. If the access record of the user 1 and the access record of the user 2 are taken as input service scene classification models, the models can output the service information of the user 1, namely the service scene identification of the access record of the user 1, and the service information of the user 2, namely the service scene identification of the access record of the user 2.

It should be noted that, the structure of the conventional neural network model is relatively simple, as shown in fig. 5 (a), including: an input layer, a hidden layer, and an output layer. And inputting data at an input layer, training the hidden layer, and outputting a result at an output layer. The RNN model differs from the conventional neural network model in that the result of each training returns to the hidden layer for training, and the structure of the RNN model is shown in fig. 5 (b). The RNN model is used for training because it is effective in processing sequence data (access records are time-ordered sequence data).

For example, as explained in connection with the above example, if a user transacts a money transfer transaction, which requires access to pages A-B-C, and transacts a redemption fund transaction, which requires access to pages A-F-G. There are cases where the user mainly wants to perform a transfer transaction operation, but the amount of money on the user's account is insufficient, so that the user opens a fund-related page to redeem a fund to the account, and then the transfer transaction operation is performed, i.e., the user's visit is recorded as page a-page F-page G-page a-page B-page C. The traditional neural network model, when trained, splits the visit record into two parts for recognition, page a-page F-page G, as redemption funds and page a-page B-page C, as transfer transactions. That is, the analysis result of the conventional neural network model is that the user transacts two businesses, i.e., makes two transactions, redeems funds and transfers. But the user mainly wants to make a transfer transaction, because the account is not full of money, the user additionally transacts the operation of redeeming funds on the account so as to realize the purpose of the transfer transaction. Therefore, the conventional neural network model has a disadvantage in recognizing the effect of the business that the user mainly wants to handle.

The RNN model can effectively solve the above problems because the RNN model has a good processing effect on sequence data. For example, corresponding to the visit record page A-page F-page G-page A-page B-page C of the user, the RNN model identifies the page A-page F-page G as a redemption fund in the training process, the result of the time is continuously input into the hidden layer for training, namely the user directly completes the operation of the transfer transaction after redeeming the fund, the RNN model considers that the redemption fund service is prepared for the transfer transaction service, and then the main service corresponding to the visit record page A-page F-page G-page A-page B-page C is determined to be the transfer transaction. Compared with the traditional neural network model, the RNN model does not simply divide the training data into each part for recognition during training, but continues to substitute the recognition result of the front part into the next recognition process, and performs training in consideration of the context of the training data, so that the RNN model has a better effect in processing data with a sequential relationship (such as a time sequential relationship or a logical sequential relationship).

That is, the RNN model includes an input layer, a hidden layer, and an output layer. The training data set includes a plurality of training data. Inputting the training data set into an RNN model for training to obtain a service scene classification model, which specifically comprises: inputting training data into an input layer of the RNN model, inputting training results (such as parameters of the RNN model obtained by the training) into the hidden layer again after the training of the hidden layer so as to participate in the next training of the training data in the hidden layer, and finally outputting output results corresponding to the training data by an output layer (the output results are service information corresponding to the training data).

After the satisfaction degree collection device obtains the business scene classification model, the business scene classification model can be stored in the satisfaction degree collection device for use by the satisfaction degree collection device. For example, after the satisfaction collecting device obtains the service scene classification model, the service transacted by the user can be determined according to the access record of the user in the first log data of the user, and then the satisfaction of the corresponding user under different services can be determined according to the service transacted by the user and the second log data (the second log data includes information about whether the service transacted by the user is successful), so as to realize the function of collecting the user satisfaction. That is, the above-described "satisfaction collection procedure" is executed. The "satisfaction collection procedure" may include S303 to S306 as follows.

S303, acquiring first log data of a plurality of users.

Wherein the first log data may include: and the access record and the user identification, wherein the access record can comprise a plurality of page URLs with time sequence generated when the user accesses the target application within a preset time period by using the terminal.

The satisfaction collecting means may acquire log data of a plurality of users.

In some embodiments, the satisfaction collection means may access the user's terminal, and obtain the user's log data from the user's terminal. The log data is data describing user identification, access record and the like generated by the terminal when the user accesses the target application by using the terminal. The target application may be a bank-like application, such as cell phone banking. Or other applications that support transacting business.

For example, if the satisfaction degree of the user in the mobile phone bank within the predetermined time period is to be collected, the satisfaction degree collecting device may access the terminal of the user and obtain the log data generated by the terminal within the predetermined time period from the terminal. The predetermined period of time may be a predefined span of time according to demand.

For example, in the case where it is necessary to collect the user satisfaction of business handling at a mobile banking station of yesterday, the satisfaction collecting means may access terminals of a plurality of users who have handled business using the mobile banking station, and acquire log data generated at the terminals of the users of yesterday from the terminals. For example, two pieces of log data are collected, respectively: log data of user 1 and log data of user 2. The log data of user 1 includes: user 1, at 9:00-9: page A-page B-page C are visited in sequence one or more times in the visit period of 10; user 1, at 12:00-12: page a-page D-page E are accessed sequentially one or more times within the access period of 10. The log data of the user 2 includes: user 2, at 9:00-9: page a-page F-page G are accessed sequentially one or more times within the access period of 10.

S304, determining the transaction information of the user according to the first log data of the user and the service scene classification model, and acquiring second log data of the user according to the identification of the user and the transaction information of the user.

The satisfaction degree collecting device can also obtain information whether each business transacted by each user is successful or not, namely obtain second log data of the user after obtaining the first log data of a plurality of users, namely obtaining the access records of each user and the identification of the user.

In some embodiments, after the first log data can be acquired by the satisfaction collecting device, the first log data is input into the service scene classification model trained in advance to determine the transaction information corresponding to the first log data.

Illustratively, as described in the foregoing embodiments, the business scenario classification model has a function of determining a business to handle based on the access record. Therefore, the satisfaction degree collecting device can input the first log data of each user into the business scene classification model, and can determine one or more businesses handled by the user corresponding to each first log data. The page URLs in the access record have a certain time sequence, and the access time of the page URLs is recorded in the access record, so that the satisfaction degree collecting device can also determine the access time period corresponding to each service. Based on the determined business handled by the user, the access time period corresponding to each business, and the user identification, the satisfaction degree collecting device can also establish a transaction behavior report for recording which user handles which business in which time period.

For example, taking user 1 as an example in the above embodiment, the satisfaction collecting device inputs the log data of user 1 into the service scene classification model, that is, user 1 is in 9:00-9: page a-page B-page C are accessed one or more times in sequence within an access period of 10, at 12:00-12:10, sequentially accessing log data such as a page A, a page D and a page E one or more times in an access time period, inputting a service scene classification model, and obtaining services transacted by a user, wherein the service scene classification model comprises the following steps: transfer transaction and fund purchase, and the corresponding access time period of the transfer transaction is 9:00-9:10, the purchase fund corresponds to an access period of 12:00-12: 10.

the satisfaction collecting device can establish a transaction behavior report based on the obtained business, the access time interval corresponding to each business and the user identification. That is, the transaction behavior report includes the user's identification, the business, and the access period. As an example, the transaction behavior report is shown in Table 1 below:

TABLE 1

From table 1, the service transacted by the user and the time period (i.e., access time period) corresponding to transacting the service can be obtained. If the user 1 transacts the transfer transaction, the corresponding access period is 9:00-9: 10; transacting the purchase fund service, wherein the corresponding access time period is 12:00-12: 10.

further, after acquiring the transaction information such as the service transacted by each user and the access time period corresponding to each service, the satisfaction collecting device may acquire second log data of the user according to the user identifier and the transaction information of the user, where the second log data includes information on whether the service transacted by the user is successful in the access time period of the service.

It should be noted that, when the user performs a transaction operation, no matter whether the transaction operation is successful or not, corresponding information is recorded in the log data of the back-end device, for example, the back-end device generates a transaction code and stores the transaction code in the log data every time the user successfully completes a transaction, and if the user fails to complete a transaction, the back-end device does not generate the transaction code, and a reason for error reporting is generated. Therefore, corresponding log data can be obtained from the back-end log to obtain whether the user successfully completes the transaction.

For example, after acquiring one or more services transacted by each user and the access time period corresponding to each service, the satisfaction collecting apparatus may access log data stored in the backend device, and acquire second log data at each user from the log data according to the user identifier and the access time period of the service, where the second log data includes: the information of whether the business transacted by the user is successful in the access time period of each business transacted by the user includes the triggering times, the transaction code when the business is successful and the reason of the transaction failure when the business is failed.

For example, continuing with the example of user 1 in the above embodiment, user 1 is at 9:00-9:10 transacts the account transfer transaction, the satisfaction collecting device accesses the log data in the back-end device, and according to the user identifier (such as the user 1) and the access time period (such as 9:00-9: 10) corresponding to the transaction, the second log data of the user 1 can be obtained, wherein the second log data comprises the log data of the user 1 in the access time period 9:00-9: and (10) processing information whether the transaction is successful or not. Such as: the second log data of the user 1 includes: user 1 is at 9:00-9:10, total 10 transactions are triggered, wherein 7 transactions are successful, 7 transaction codes are generated, 3 transactions are unsuccessful, and 3 transaction failure reasons are generated, namely 2 times of password input errors and 1 time of verification code input errors. Similarly, user 1 is at 12:00-12: the transaction of purchasing funds is transacted 10, and the second log data of the user 1, which is also obtained by the satisfaction collecting means, further includes: user 1 during visit period 12:00-12: the information about whether the transaction is successful or not in 10 includes: user 1 is at 12:00-12:10, a total of 5 transactions are triggered, 4 of which are successful and 4 transaction codes are generated, 1 is unsuccessful and 1 transaction failure reason is generated, which is response timeout.

S305, determining the user satisfaction degree of the target service according to the transaction information and the second log data of each user in the plurality of users.

After the transaction information and the second log data of each of the plurality of users are acquired, the satisfaction degree collecting device may determine the user satisfaction degree of the target service according to the transaction information and the second log data of each of the plurality of users. The concrete steps include the following S305a-S305 b.

S305a, for each user in the plurality of users, determining the transaction success rate of each service transacted by the user through the target application within a preset time period according to the second log data of the user.

And after the satisfaction degree collecting device obtains the second log data of the user, the transaction success rate of each service transacted by the user can be determined.

In some embodiments, after acquiring the second log data of the user, the satisfaction collecting device determines the transaction success rate of each service transacted by the user according to the information about whether the user transaction is successful in the second log data.

For example, the satisfaction collecting means may, based on the formula, according to the number of transaction codes in the second log data and the number of occurrences of the transaction failure reason: and the transaction success rate is the number of the transaction codes/(the number of the transaction codes + the number of the transaction failure reasons), so as to determine the transaction success rate. In addition, the satisfaction collecting device can also establish a background transaction statistical report according to the transaction information (or transaction behavior report) of each user, the second log data and the transaction success rate, and the method comprises the following steps: user identification, triggering times, service, transaction success rate and transaction failure reason.

For example, continuing with the example of the user 1 in the above embodiment, the satisfaction collecting means may, based on the number of transaction codes and the number of occurrences of the transaction failure reasons in the access period of 9:00-9:10 in the second log data of the user 1, for example, if the user 1 triggers 10 transactions in total in 9:00-9:10, the generated transaction codes are 7, and the number of occurrences of the transaction failure reasons is 3, including: the password is input incorrectly for 1 time, the authentication code is input incorrectly for 2 times, and the transaction success rate of the user 1 in the access period 9:00-9:10 is determined to be 7/(7+3) ═ 70%. Similarly, if the user 1 has 4 transaction codes in the access period 12:00-12:10, and the reason for the failure of the transaction is that the response is overtime for 1 time, the satisfaction collecting means may determine that the transaction success rate of the user 2 in the time period 12:00-12:10 is 80%. In addition, as described in the previous embodiment, the transaction information of the user 1 includes that the user 1 transacts the transfer transaction, and the corresponding access time period is 9:00-9: 10; transacting the purchase fund service, wherein the corresponding access time period is 12:00-12:10 such information. Therefore, the satisfaction collecting means may determine that the transaction success rate of the user 1 transacting the transfer transaction is 70% and the transaction success rate of the purchase fund transaction is 80% based on the aforementioned determination result.

In addition, the satisfaction collecting device may establish a background transaction statistical report based on the transaction information (or a transaction behavior report including the user identifier, the service transacted by the user, and the access time period of each service) of each user, the second log data (including the number of triggers, the transaction code when successful, and the transaction failure reason when failed) and the transaction success rate of each service, as shown in table 2 below:

TABLE 2

The services handled by the user, the transaction success rate of each service and the reason for the failure of the transaction can be obtained from table 2. If the user 1 transacts the transfer service, the transaction success rate of the service is 70%, and the transaction failure reasons are password input errors and verification code input errors. The user 1 also transacts the purchase fund service, the transaction success rate of the service is 80%, and the reason for the transaction failure is response overtime.

Similarly, for other users, the satisfaction degree collecting device can also obtain each service transacted by using the target application and the transaction success rate of each service. The transaction failure reason of each service can also be obtained.

S305b, determining the user satisfaction of the target service according to the service transacted by the target application in the preset time period in the transaction information of the plurality of users and the transaction success rate of each service transacted by each user in the preset time period.

The satisfaction degree collecting device can determine the user satisfaction degree of the target service after acquiring the transacted services of a plurality of users and the transaction success rate of each service. Wherein, the transaction success rate of the business transaction is in direct proportion to the user satisfaction. The target service includes all services handled by the plurality of users.

For example, for each service in the target service, the satisfaction collecting means may determine the average transaction success rate of the service according to the transaction success rates of a plurality of users handling the service. And the satisfaction collecting device determines the user satisfaction of the corresponding service according to the average transaction success rate of the service.

For example, in combination with the above embodiment, if the user 1 transacts the transfer transaction service, triggers 10 transactions, the transaction success rate is 70%, and the user 2 transacts the transfer transaction service, triggers 5 transactions, and the transaction success rate is 80%, the average transaction success rate of the transfer transaction service can be determined by using the following formula:

that is, the average transaction success rate of the transfer transaction service is 73%, and thus the satisfaction collecting means may determine that the user satisfaction of the transfer transaction service is 73%. Similarly, if the user 1 transacts the fund purchasing service and triggers 5 transactions with a transaction success rate of 80%, and if the user 2 transacts the fund purchasing service and triggers 20 transactions with a transaction success rate of 60%, the satisfaction degree collecting device may determine the average transaction success rate of the fund purchasing service, so as to determine the user satisfaction degree of the fund purchasing service, for example, 64%.

In addition, the satisfaction collecting device can also establish a service satisfaction report according to the target service and the triggering times of each service, namely the service satisfaction report comprises the service, the triggering times and the user satisfaction. For example, in conjunction with the above example, the satisfaction collecting device creates a business satisfaction report, as shown in table 3 below:

TABLE 3

The service handled in a preset time period, the transaction triggering times of each service and the user satisfaction can be obtained from the table 3, for example, the transfer transaction service is handled in one day of yesterday, the transaction is triggered for 15 times, and the user satisfaction is 73%; a purchase fund service was also transacted, which triggered the transaction 25 times with a user satisfaction of 64%.

S306, determining the dissatisfaction reason of the target service user according to the transaction information and the second log data of each user in the plurality of users.

The satisfaction degree collecting device can also determine the dissatisfaction reason of the target service according to the transaction information of each user and the second log data.

In some embodiments, the satisfaction collecting device may further determine, according to the services transacted by the user and one or more transaction failure reasons corresponding to each service, the transaction failure reason N before the occurrence number as the user dissatisfaction reason of the service. N is an integer greater than or equal to 1.

For example, in combination with the example in the above embodiment, taking N as 1 as an example, after analyzing the transaction information and the second log data of a plurality of users within a predetermined time period, 4 transaction failure reasons in the transfer transaction service are obtained, where a password is input incorrectly for 1 time, an authentication code is input incorrectly for 2 times, and a response is overtime for 1 time, the satisfaction degree collecting device may determine that the password input error is a reason for dissatisfaction of the user handling the transfer transaction service. Similarly, in the transaction failure reasons for purchasing the fund service obtained by analyzing the transaction information and the second log data of the plurality of users within the predetermined time period, if the password is input incorrectly for 1 time and the response is overtime for 8 times, the satisfaction degree collecting device may determine the response overtime as the dissatisfaction reason for the user transacting the fund purchasing service.

In another example, the satisfaction collecting means may determine, as the user dissatisfaction reason for the service, a transaction failure reason for which the number of occurrences of the transaction failure reason under each service is greater than a threshold. The threshold value may be preset, and the size of the threshold value may be determined according to actual conditions.

In addition, after determining the dissatisfaction reason of the user of each service, the satisfaction degree collecting device can also add the following factors to the service satisfaction degree report form: the reason why the user is not satisfied. As shown in table 4 below:

TABLE 4

Business	Number of triggers	Degree of satisfaction of user	Reasons for dissatisfaction of users
				Transfer transactions	15	73％	Verification code entry errors
Purchasing funds	25	64％	Response timeout

From table 4, the user satisfaction and the reason for the user dissatisfaction of each service can be obtained, such as: the user satisfaction degree of the transfer transaction service is 73%, and the reason for the user dissatisfaction is that the verification code is input wrongly; the customer satisfaction with the purchase of the fund service is 64%, and the reason for the customer dissatisfaction is the response timeout.

It can be understood that the user satisfaction and the user dissatisfaction reason of each service are determined, and for the service with lower user satisfaction, the service strategy can be adjusted according to the corresponding user dissatisfaction reason to optimize the system. For example, the user of the transfer transaction service is dissatisfied because the verification code is input incorrectly, and the verification code sent to the user terminal by the system can be adjusted from 6 bits to 4 bits when the user transacts the transfer service, so that the frequency of incorrect input of the user is reduced. If the dissatisfaction reason of the fund purchase is response timeout, the system can properly prolong the response time when the user transacts the fund purchase service, and the user can operate in more time.

Illustratively, the above embodiment will be briefly summarized below with reference to fig. 6. When a user accesses a front-end page of an application (such as a mobile banking) capable of handling business on a mobile phone terminal, a front-end log record is generated in the mobile phone terminal to record the access behavior of the user. When the user satisfaction needs to be collected, the satisfaction collecting device can access a front-end log in a mobile phone terminal of the user to obtain log data, wherein the log data comprises a user identifier and an access record. The satisfaction degree collecting device can input the user identification and the access record into the RNN model to classify the service scene so as to obtain a transaction behavior report. Moreover, the user transacts the service also generates a back-end log in the back-end server, and the back-end log records whether the user transacts the service successfully or not. The satisfaction degree collecting device can access the back-end log to obtain log data so as to obtain a background transaction statistical report. The satisfaction collecting device can obtain a service satisfaction report according to the transaction behavior report and the background transaction statistical report. The satisfaction and the use experience of the user in handling each service can be obtained from the service satisfaction report, and the system is optimally designed according to the experience of the user.

Furthermore, the method adopts the RNN model to determine the business which the user wants to handle, the RNN model has good effect of processing the sequence data, the obtained data is more accurate, and the collected user satisfaction degree is more reliable. In addition, the method can analyze the system log data by acquiring a large amount of system log data in real time, so that the user satisfaction degrees under different service scenes can be comprehensively determined in real time. Moreover, the dissatisfaction reason of the user of each service scene can be determined, and the system can be optimized according to the dissatisfaction reason of the user aiming at the service scene with low satisfaction degree, so that the system is more perfect, comfortable use experience is brought to the user, and the user can be better maintained.

In an exemplary embodiment, the present application further provides a satisfaction collection device. The satisfaction collecting means may comprise one or more functional modules for implementing the satisfaction collecting method of the above method embodiments.

For example, fig. 7 is a schematic diagram of a satisfaction collecting device according to an embodiment of the present application. As shown in fig. 7, the satisfaction collecting means includes: an obtaining module 701 and a determining module 702. The obtaining module 701 is connected with the determining module 702.

An obtaining module 701, configured to obtain first log data of multiple users; wherein the first log data includes: and the access record comprises a plurality of page URLs with time sequence generated when the user accesses the target application within a preset time period by using the terminal and the identification of the user.

For each user of a plurality of users:

the obtaining module 701 is further configured to determine transaction information of the user according to the first log data of the user and the service scene classification model, and obtain second log data of the user according to the identifier of the user and the transaction information of the user; wherein, the transaction information of the user comprises: one or more services transacted by a user through a target application in a preset time period and an access time period corresponding to each service; the second log data includes: and in the access time period corresponding to each service, the information whether each service transacted by the user through the target application is successful or not.

A determining module 702, configured to determine user satisfaction of the target service according to the transaction information and the second log data of each of the multiple users; the target service comprises all services transacted by a plurality of users through the target application within a preset time period.

In some embodiments, the determining module 702 is specifically configured to: aiming at each user in the plurality of users, determining the transaction success rate of each service transacted by the user through the target application within a preset time period according to the second log data of the user; determining the user satisfaction degree of the target service according to the service transacted by the target application in the transaction information of the plurality of users within a preset time period and the transaction success rate of each service transacted by each user within the preset time period; wherein, the success rate of service transaction is in direct proportion to the satisfaction degree of the user.

In some embodiments, the apparatus further comprises: a module 703 is established.

An establishing module 703 configured to: acquiring a training data set; the training data set comprises a plurality of pieces of training data, one piece of training data comprises an access record and corresponding service information, wherein the access record is generated when a sample user uses a terminal to access a target application to transact a service, and the access record comprises a plurality of page URLs with time sequence; the service information corresponding to the access record is self-defined; performing model training according to the training data set to obtain a business scene classification model; the service scene classification model has the function of determining the transacted service according to the access record.

In some embodiments, the determining module 702 is further configured to determine a user dissatisfaction reason of the target service according to the transaction information and the second log data of each of the plurality of users.

In some embodiments, the establishing module 703 is specifically configured to input the training data set into the RNN model for training, so as to obtain the service scene classification model.

In other embodiments, as shown in fig. 8, there is also provided a schematic diagram of a satisfaction collection apparatus, which may include: the system comprises a front-end log acquisition module, a rear-end log reading and analyzing module, an RNN service scene classification module and a scene satisfaction degree calculation module.

The four modules perform functions similar to the functions of the corresponding modules in the apparatus shown in fig. 7 described above. For example, the front-end log reading module and the back-end log reading and analyzing module have similar functions as the obtaining module 701, the scene satisfaction calculating module has similar functions as the determining module 702, and the RNN service scene classifying module has similar functions as the establishing module 703.

In an exemplary embodiment, the present application further provides an electronic device, which may be the satisfaction degree collecting apparatus in the foregoing method embodiment. Fig. 9 is a schematic structural diagram of a satisfaction collecting device according to an embodiment of the present application. As shown in fig. 9, the satisfaction collecting means may include: a processor 901 and a memory 902; memory 902 stores instructions executable by processor 901; the processor 901 is configured to execute the instructions, so that the electronic device or the network device or the manager implements the method as described in the foregoing method embodiments.

In an exemplary embodiment, the present application further provides a computer-readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a computer, cause the computer to implement a method as described in the preceding embodiments. The computer may be an electronic device or a network device or manager. The computer readable storage medium may be a non-transitory computer readable storage medium, for example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

In an exemplary embodiment, the present application further provides a computer program product, which when running on a computer, causes the computer to execute the above related method steps to implement the satisfaction degree collection method in the above embodiments.

The above is only an embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions within the technical scope of the present disclosure should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A satisfaction collection method, characterized in that the method comprises:

acquiring first log data of a plurality of users; wherein the first log data includes: the method comprises the steps that an access record and a user identification are obtained, wherein the access record comprises a plurality of page Uniform Resource Locators (URLs) with time sequence, which are generated when the user accesses a target application within a preset time period by using a terminal;

for each user of the plurality of users: determining transaction information of the user according to the first log data of the user and a service scene classification model, and acquiring second log data of the user according to the identification of the user and the transaction information of the user; wherein the transaction information of the user comprises: one or more services transacted by the user through the target application in the preset time period and an access time period corresponding to each service; the second log data includes: information whether each service transacted by the user through the target application is successful or not in an access time period corresponding to each service;

determining user satisfaction of a target service according to the transaction information and the second log data of each user in the plurality of users; and the target service comprises all services transacted by the plurality of users through the target application within the preset time period.

2. The method of claim 1, wherein determining user satisfaction for a target service based on the transaction information and the second log data for each of the plurality of users comprises:

for each user in the plurality of users, determining the transaction success rate of each service transacted by the user through the target application within the preset time period according to the second log data of the user;

determining the user satisfaction degree of the target service according to the service transacted by the target application in the preset time period in the transaction information of the users and the transaction success rate of each service transacted by each user in the preset time period through the target application;

wherein, the transaction success rate of the business transaction is in direct proportion to the user satisfaction.

3. The method according to claim 1 or 2, characterized in that the method further comprises:

acquiring a training data set; the training data set comprises a plurality of pieces of training data, one piece of training data comprises an access record and corresponding service information, wherein the access record is generated when a sample user accesses the target application by using a terminal to handle one service, and the access record comprises a plurality of page URLs with time sequence; the service information corresponding to the access record is self-defined;

performing model training according to the training data set to obtain the service scene classification model; the service scene classification model has the function of determining the transacted service according to the access record.

4. The method according to claim 1 or 2, wherein in case of failure of a business transaction transacted by a user through the target application, the second log data further comprises: a transaction failure reason;

the method further comprises the following steps:

and determining the dissatisfaction reason of the target service according to the transaction information and the second log data of each user in the plurality of users.

5. The method of claim 3, wherein the model training according to the training data set to obtain the traffic scenario classification model comprises:

and inputting the training data set into a Recurrent Neural Network (RNN) model for training so as to obtain the service scene classification model.

6. A satisfaction-gathering device, characterized in that said device comprises: the device comprises an acquisition module and a determination module;

the acquisition module is used for acquiring first log data of a plurality of users; wherein the first log data includes: the method comprises the steps that an access record and a user identification are obtained, wherein the access record comprises a plurality of page Uniform Resource Locators (URLs) with time sequence, which are generated when the user accesses a target application within a preset time period by using a terminal;

for each user of the plurality of users: the acquisition module is further configured to determine transaction information of the user according to the first log data of the user and a service scene classification model, and acquire second log data of the user according to the identifier of the user and the transaction information of the user; wherein the transaction information of the user comprises: one or more services transacted by the user through the target application in the preset time period and an access time period corresponding to each service; the second log data includes: information whether each service transacted by the user through the target application is successful or not in an access time period corresponding to each service;

the determining module is used for determining the user satisfaction degree of the target service according to the transaction information and the second log data of each user in the plurality of users; and the target service comprises all services transacted by the plurality of users through the target application within the preset time period.

7. The apparatus of claim 6,

the determining module is specifically configured to:

8. The apparatus of claim 6 or 7, further comprising: establishing a module;

the establishing module is configured to:

9. An electronic device, characterized in that the electronic device comprises: a processor and a memory;

the memory stores instructions executable by the processor;

the processor is configured to, when executing the instructions, cause the electronic device to implement the method of any of claims 1-5.

10. A computer-readable storage medium, the computer-readable storage medium comprising: computer software instructions;

the computer software instructions, when executed in an electronic device, cause the electronic device to implement the method of any of claims 1-5.