A user satisfaction analysis method and system based on user event training

Abstract

The present invention provides a user satisfaction analysis method and system based on user event training, including: analyzing event results of activity events generated by users when operating a Web interface, generating a user activity log of the user, identifying immediate response information and delayed response information generated by the Web interface, inputting the immediate response information and delayed response information into the user activity log for event matching, obtaining a plurality of activity responses of the Web interface and the number of response repetitions corresponding to each of the activity responses, searching for repeated activity events in the activity log, using a neural network to perform annotation training on the repeated activity events, determining potential defects of the Web interface, generating and displaying a user satisfaction report, performing sentiment analysis on the user log, judging the user's usage experience, and thus predicting the user's satisfaction with the Web.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a user satisfaction analysis method and system based on user event training. Background Art

[0002] Customer satisfaction reflects the value of an enterprise. It has a very important guiding significance for the marketing of an enterprise, both at the strategic level and at the tactical level. The investigation of customer satisfaction and the connection of the results of the investigation with the attributes of products / services have become a tool for market-driven quality methods. On the other hand, with the development of information technology, a huge amount of information has been accumulated in the huge database of telecom operators. How to apply data mining technology to all aspects of operation, management and decision-making, and obtain effective information from massive data can provide inspiring guidance for enterprises. Therefore, how to effectively and accurately evaluate user satisfaction and use data mining methods to perform correlation analysis on massive data is an urgent need of the communications industry. Web page usage mining is a sub-field of data mining. According to the type of data mined, it is usually divided into web service log data, application service log data, and application-level log data. Application-level log data contains various new types of user events that can be monitored in applications. However, this type of monitoring method can only perform superficial analysis, or perform corresponding analysis based on the opinions put forward by users. The analysis results are not comprehensive and authoritative.

[0003] Therefore, the present invention provides a method and system for analyzing user satisfaction based on user event training. Summary of the invention

[0004] The present invention provides a user satisfaction analysis method and system based on user event training, which performs sentiment analysis on user logs, judges user usage experience, and thus predicts user satisfaction with the Web.

[0005] The present invention provides a user satisfaction analysis method based on user event training, comprising:

[0006] Step 1: Collect activity events generated by users when operating on the Web interface, analyze the event results corresponding to each activity event, and generate a user activity log of the user;

[0007] Step 2: Identify in the response log of the Web interface the immediate response information and the delayed response information generated by the Web interface in response to the user operation;

[0008] Step 3: Input the immediate response information and the delayed response information into the user activity log for event matching, and obtain a number of activity responses of the Web interface and the number of response repetitions corresponding to each activity response;

[0009] Step 4: Search for repeated activity events in the activity log, use a neural network to perform annotation training on the repeated activity events, determine potential defects of the Web interface, generate a user satisfaction report on the Web interface, and display it.

[0010] In one practicable manner,

[0011] The step 1 comprises:

[0012] Step 11: Collecting activity data generated when the user performs operation activities on the Web interface, establishing the user's access track according to the activity data, identifying the encryption level corresponding to each track point in the access track, and encrypting the corresponding data segment in the activity data according to the encryption level to obtain a plurality of encrypted data segments;

[0013] Step 12: Establishing the activity event corresponding to the user according to the encrypted data segment, identifying the event start point and event end point corresponding to each activity event on the Web interface, arranging the activity events according to the event start point and event end point, and generating the event chain of the user;

[0014] Step 13: Analyze the event logical relationship between different activity events in the event chain, determine the event result corresponding to each activity event in combination with the event start point and event end point corresponding to each activity event, and obtain the operation content and operation time corresponding to each event result to generate the user activity log of the user.

[0015] In one practicable manner,

[0016] The step 2 comprises:

[0017] Step 21: determining the operation time period of the user on the Web interface according to the user activity log, obtaining response data of the Web interface within the operation time period, generating a response log of the Web interface, and identifying the response information corresponding to each event result in the response log;

[0018] Step 22: performing a first classification on the response information according to the operation action corresponding to each of the response information to generate a first response classification group, and performing a second classification on the response information according to the operation element corresponding to each of the response information to generate a second response classification group;

[0019] Step 23: performing action training on each of the first response classification groups to obtain the corresponding operation pause preference when the user performs the corresponding operation action, and performing element recognition on each of the second response classification groups to obtain the basic response duration corresponding to each execution element;

[0020] Step 24: respectively obtain the operation pause preference and basic response time corresponding to each of the response information, generate an immediate response range corresponding to the response information, analyze in the response log whether the actual response time of each of the response information is within the corresponding immediate response range, and divide the response information into immediate response information and delayed response information according to the analysis result.

[0021] In one practicable manner,

[0022] The step 3 comprises:

[0023] Step 31: extracting a first response key corresponding to each of the immediate response messages, and extracting a second response key corresponding to each of the delayed response messages, using the first response key to identify a first keyword of the response in the user activity log, and using the second response key to identify a second keyword of the response in the user activity log;

[0024] Step 32: Determine a first activity event corresponding to each instant response message according to a matching relationship between the first response key and the first keyword, determine a second activity event corresponding to each instant response message according to a matching relationship between the second response key and the second keyword, establish a first response purpose according to the first activity event, and establish a second response purpose according to the second activity event;

[0025] Step 33: identifying a first response information point corresponding to the first response destination in the immediate response information, treating the first response information point as the immediate response information end point of the immediate response information, generating an immediate active response of the Web interface, identifying a second response information point corresponding to the second response destination in the delayed response information, intercepting the information segment before the second response information point in the delayed response information, and generating a delayed active response of the Web interface;

[0026] Step 34: The immediate activity response and the delayed activity response are regarded as activity responses, response attributes corresponding to each of the activity responses are obtained respectively, activity responses with consistent response attributes are regarded as repeated responses, and the number of response repetitions corresponding to each of the repeated responses is counted respectively.

[0027] In one practicable manner,

[0028] Also includes:

[0029] Identifying in each of the instant response information the first response information point corresponding to the first response destination;

[0030] Acquire, in the instant response information, a first information time difference between the first response information point and the corresponding instant response information end point and a second information time difference between the first response information point and the corresponding instant response information start point;

[0031] When the time difference of the first information is greater than the time difference of the second information, the corresponding immediate response information is regarded as delayed response information;

[0032] When the second information time difference is greater than the preset time difference, the corresponding immediate response information is regarded as delayed response information.

[0033] In one practicable manner,

[0034] The step 4 comprises:

[0035] Step 41: determining the number of activity repetitions of the repeated activity event contained in the activity log according to the number of response repetitions corresponding to the activity response, establishing an event statistics table based on the number of activity repetitions, identifying the repeated activity events contained in the activity log, and filling the repeated activity events in the event statistics table according to the activity time corresponding to each repeated activity event;

[0036] Step 42: Analyze the activity time distribution information corresponding to each repeated activity event in the event statistics table, classify and annotate the event in the event statistics table, generate the annotated event table of the user, input the annotated event table into a preset neural network for deep training, obtain the valid activities and invalid activities contained in the distribution information of each activity event, and mark the effects in the corresponding repeated activity events respectively;

[0037] Step 43: Count the invalid information corresponding to each of the repeated activity events, determine the potential defects of the Web interface according to the invalid range corresponding to each of the invalid information, trace each of the potential defects respectively, obtain the historical defect remediation information of the Web interface, generate the user satisfaction report on the Web interface according to the potential defects and the corresponding historical defect remediation information, and display it.

[0038] In one practicable manner,

[0039] Also includes:

[0040] Acquire the suggestion information submitted by the user, and extract the user's usage defects from the suggestion information;

[0041] A defect list is generated and displayed according to the usage defects and potential defects of the Web interface.

[0042] In one practicable manner,

[0043] Also includes:

[0044] Obtain satisfaction reports corresponding to a number of users, perform fusion analysis on the satisfaction reports, obtain the number of presentations corresponding to each of the potential defects, and generate a treatment priority corresponding to each of the potential defects.

[0045] The present invention provides a user satisfaction analysis system based on user event training, comprising:

[0046] An event collection module is used to collect activity events generated by users when operating on the Web interface, analyze the event results corresponding to each activity event, and generate a user activity log for the user;

[0047] A response analysis module, used to identify in the response log of the Web interface the immediate response information and the delayed response information generated by the Web interface in response to the user operation;

[0048] An abnormality identification module, used for inputting the immediate response information and the delayed response information into the user activity log for event matching, and obtaining a plurality of activity responses of the Web interface and a response repetition number corresponding to each activity response;

[0049] The defect determination module is used to search for repeated activity events in the activity log, use a neural network to perform annotation training on the repeated activity events, determine potential defects of the Web interface, generate a user satisfaction report on the Web interface and display it.

[0050] In one practicable manner,

[0051] The defect determination module comprises:

[0052] a preprocessing unit, configured to determine the number of activity repetitions of the repeated activity events contained in the activity log according to the number of response repetitions corresponding to the activity response, establish an event statistics table based on the number of activity repetitions, identify the repeated activity events contained in the activity log, and fill the repeated activity events in the event statistics table according to the activity time corresponding to each repeated activity event;

[0053] An event marking unit is used to analyze the activity time distribution information corresponding to each repeated activity event in the event statistics table, classify and mark in the event statistics table, generate a marked event table for the user, input the marked event table into a preset neural network for deep training, obtain the valid activities and invalid activities contained in the distribution information of each activity event, and mark the effects in the corresponding repeated activity events respectively;

[0054] A deep analysis unit is used to count the invalid information corresponding to each of the repeated activity events, determine the potential defects of the Web interface according to the invalid range corresponding to each of the invalid information, trace the source of each of the potential defects respectively, obtain the historical defect remediation information of the Web interface, generate the user satisfaction report on the Web interface according to the potential defects and the corresponding historical defect remediation information, and display it.

[0055] The achievable beneficial effects of the above technical solution are: in order to objectively analyze the user's experience of using the Web system and improve the user's interest in use, the user's activity events on the Web interface are first collected, and the user's activity log is established by analyzing the event results corresponding to each activity event. Then, the immediate response information and delayed response information of the Web interface are identified in the response log of the Web interface, and the corresponding activity events are further matched for the immediate response information and the delayed response information, and repeated activity events are screened. Finally, the repeated activity events are labeled and trained using a neural network to find defects in the Web interface and generate a user satisfaction report on the Web interface. In this way, synchronous analysis can be performed when the user uses the Web interface, which solves the limitation of relying on questionnaires to analyze user satisfaction in the past. By summarizing the user's usage experience at all times, the defects of the system can be better determined so that adjustments can be made to improve the fluency of the system and improve user satisfaction.

[0056] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0057] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0059] Figure 1 A schematic diagram of the workflow of a user satisfaction analysis method based on user event training in an embodiment of the present invention;

[0060] Figure 2 The present invention is a schematic diagram of the composition of a user satisfaction analysis system based on user event training in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0062] Example 1

[0063] This embodiment provides a user satisfaction analysis method and system based on user event training, such as Figure 1 As shown, including:

[0064] Step 1: Collect activity events generated by users when operating on the Web interface, analyze the event results corresponding to each activity event, and generate a user activity log of the user;

[0065] Step 2: Identify in the response log of the Web interface the immediate response information and the delayed response information generated by the Web interface in response to the user operation;

[0066] Step 3: Input the immediate response information and the delayed response information into the user activity log for event matching, and obtain a number of activity responses of the Web interface and the number of response repetitions corresponding to each activity response;

[0067] Step 4: Search for repeated activity events in the activity log, use a neural network to perform annotation training on the repeated activity events, determine potential defects of the Web interface, generate a user satisfaction report on the Web interface, and display it.

[0068] In this example, the activity event represents the event generated when the user performs activities in the Web network;

[0069] In this example, the event result indicates the result generated after the user operates the web interface;

[0070] In this example, the immediate response information indicates that the Web system responds to the user's operation in a short time, and the delayed response information indicates that the Web system does not respond to the user's operation in a short time, but responds after a long time.

[0071] In this example, both the immediate response information and the delayed response information may correspond to an activity event;

[0072] In this example, the neural network is initially manually annotated with data to record users with different proficiency levels using different functional modules of multiple systems (distinguished by primary routing) to generate event-time series. At the same time, the user can record whether he is satisfied with the page operation at any time based on the current time. When the user uses the system, the network predicts the current user satisfaction in real time. When the satisfaction exceeds (satisfied) or is lower than (threshold), a prompt pops up to prompt the user to choose whether he is satisfied with the current page. If the user chooses, the result of the current user operation sequence is recorded to generate an online data set, collect the online data set regularly, train the model twice, and iterate the above operations until the network converges to the optimal state.

[0073] The working principle and beneficial effects of the above technical solution are as follows: In order to objectively analyze the user's experience of using the Web system and improve the user's interest in using the system, the user's activity events on the Web interface are first collected, and the user's activity log is established by analyzing the event results corresponding to each activity event. Then, the immediate response information and delayed response information of the Web interface are identified in the response log of the Web interface, and the corresponding activity events are further matched for the immediate response information and the delayed response information, and repeated activity events are screened. Finally, the repeated activity events are labeled and trained using a neural network to find defects in the Web interface and generate a user satisfaction report on the Web interface. In this way, synchronous analysis can be performed when the user uses the Web interface, which solves the limitation of relying on questionnaires to analyze user satisfaction in the past. By summarizing the user's usage experience at all times, the defects of the system can be better determined so that adjustments can be made to improve the fluency of the system and improve user satisfaction.

[0074] Example 2

[0075] On the basis of Example 1, the user satisfaction analysis method based on user event training, step 1, comprises:

[0076] Step 11: Collecting activity data generated when the user performs operation activities on the Web interface, establishing the user's access track according to the activity data, identifying the encryption level corresponding to each track point in the access track, and encrypting the corresponding data segment in the activity data according to the encryption level to obtain a plurality of encrypted data segments;

[0077] Step 12: Establishing the activity event corresponding to the user according to the encrypted data segment, identifying the event start point and event end point corresponding to each activity event on the Web interface, arranging the activity events according to the event start point and event end point, and generating the event chain of the user;

[0078] Step 13: Analyze the event logical relationship between different activity events in the event chain, determine the event result corresponding to each activity event in combination with the event start point and event end point corresponding to each activity event, and obtain the operation content and operation time corresponding to each event result to generate the user activity log of the user.

[0079] In this example, the access trace represents the trace generated by the user when performing operations to access the Web system;

[0080] In this example, the encryption level is related to the operation corresponding to the track point. The encryption levels include: weak encryption, standard encryption and strong encryption. For example, the page is encrypted with standard encryption when the user searches, and the page is encrypted with strong encryption when the user conducts transactions.

[0081] In this example, the event start point indicates the state of the active event, and the event end point indicates the state of the completed active event;

[0082] In this example, the event chain mark arranges the activity events generated by the user when performing the activity in chronological order.

[0083] The working principle and beneficial effects of the above technical solution are as follows: when a user operates on a Web page, an access track is established based on the user's activity data, and the corresponding data segment is further encrypted by identifying the encryption level corresponding to each track point to obtain the corresponding encrypted data segment. Then, the user's activity event is established, and an event chain about a series of operations performed by the user is generated by connecting the activity events end to end. Then, the event logic relationship contained in the event chain is analyzed to determine the event result corresponding to each activity event. Finally, the user's activity log is generated based on the operation content and operation time of each event result. In this way, not only can the user's activity log be generated in real time, but also corresponding encryption is performed when collecting data to avoid user privacy leakage. Satisfaction analysis is performed on the basis of protecting user privacy, creating a safe and reliable network environment for users.

[0084] Example 3

[0085] On the basis of Example 1, the user satisfaction analysis method based on user event training, step 2 includes:

[0086] Step 21: determining the operation time period of the user on the Web interface according to the user activity log, obtaining response data of the Web interface within the operation time period, generating a response log of the Web interface, and identifying the response information corresponding to each event result in the response log;

[0087] Step 22: performing a first classification on the response information according to the operation action corresponding to each of the response information to generate a first response classification group, and performing a second classification on the response information according to the operation element corresponding to each of the response information to generate a second response classification group;

[0088] Step 23: performing action training on each of the first response classification groups to obtain the corresponding operation pause preference when the user performs the corresponding operation action, and performing element recognition on each of the second response classification groups to obtain the basic response duration corresponding to each execution element;

[0089] Step 24: respectively obtain the operation pause preference and basic response time corresponding to each of the response information, generate an immediate response range corresponding to the response information, analyze in the response log whether the actual response time of each of the response information is within the corresponding immediate response range, and divide the response information into immediate response information and delayed response information according to the analysis result.

[0090] In this example, the response data refers to the data generated when the Web operation interface responds to the user's operation;

[0091] In this example, the response information indicates the response of the Web page to the event result performed by the user;

[0092] In this example, the first response classification group includes response information of a plurality of consistent operation actions, and the second response classification group includes response information of a plurality of consistent operation elements;

[0093] In this example, the operation action refers to the action performed by the user, including input, mouse click, double click, right click, scroll, route switching, page hiding, displaying, etc.;

[0094] In this example, the operation element indicates that a specific mark needs to be given to the HTML element during web application development, or a unique mark needs to be generated according to a rule. The rule is to automatically generate "type_UUID" for buttons, input boxes, elements bound to events, elements with text content, images, and SVG.

[0095] In this example, the operation pause preference represents the personal pause that the user generates when performing an operation;

[0096] In this example, the basic response time indicates the time required for the web interface to respond;

[0097] In this example, the immediate response range indicates the time range for an immediate response to an event, and the duration of the immediate response range corresponding to different activity times may not be the same.

[0098] The working principle and beneficial effects of the above technical solution are as follows: by analyzing the user activity log to determine the operation time period of the Web interface, the corresponding response data is selected to construct a response log, and then the response information corresponding to each event result is identified in the response log, and the response information is further classified according to the operation action and operation element corresponding to each response information, and two different classification results are obtained. Then, action training is performed on one of the classification groups to determine the user's pause preference, and element identification is performed on the other group to obtain its corresponding basic execution time. In this way, the minimum response time of the Web interface for the active event can be determined, so as to construct its immediate response range. Finally, the actual response time is analyzed in the response log to determine whether it is within its corresponding immediate response range, and the immediacy and delay of the response information are determined. In this way, the immediate response information and the delayed response information can be quickly separated, and the attribute analysis of the response information is realized, which is convenient for better analysis of user satisfaction.

[0099] Example 4

[0100] On the basis of Example 1, the user satisfaction analysis method based on user event training, step 3 includes:

[0101] Step 31: extracting a first response key corresponding to each of the immediate response messages, and extracting a second response key corresponding to each of the delayed response messages, using the first response key to identify a first keyword of the response in the user activity log, and using the second response key to identify a second keyword of the response in the user activity log;

[0102] Step 32: Determine a first activity event corresponding to each instant response message according to a matching relationship between the first response key and the first keyword, determine a second activity event corresponding to each instant response message according to a matching relationship between the second response key and the second keyword, establish a first response purpose according to the first activity event, and establish a second response purpose according to the second activity event;

[0103] Step 33: identifying a first response information point corresponding to the first response destination in the immediate response information, treating the first response information point as the immediate response information end point of the immediate response information, generating an immediate active response of the Web interface, identifying a second response information point corresponding to the second response destination in the delayed response information, intercepting the information segment before the second response information point in the delayed response information, and generating a delayed active response of the Web interface;

[0104] Step 34: The immediate activity response and the delayed activity response are regarded as activity responses, response attributes corresponding to each of the activity responses are obtained respectively, activity responses with consistent response attributes are regarded as repeated responses, and the number of response repetitions corresponding to each of the repeated responses is counted respectively.

[0105] In this example, the first response key indicates a response keyword included in the immediate response information, and the second response key indicates a response keyword included in the delayed response information;

[0106] In this example, the first response information point indicates the representation form of the first response purpose in the immediate response information, and the second response information point indicates the representation form of the second response purpose in the delayed response information;

[0107] In this example, the purpose of regarding the first response information point as the immediate response information end point of the immediate response information is to: remove redundant information in the immediate response information;

[0108] In this example, the response attributes consistently represent active responses having the same response purpose and response operation.

[0109] The working principle and beneficial effects of the above technical solution are as follows: when repeated responses appear in a Web page, it means that the user has performed multiple repeated operations in a short period of time, that is, the Web system may be abnormal. In order to screen repeated activity events from multiple activity events, the immediate response information and the delayed response information are first matched with the activity events, and the response purpose of each activity event is established. Then, the immediate response information and the delayed response information are adjusted to obtain the corresponding activity responses. Finally, the activity responses with consistent response attributes are regarded as repeated responses, and the response times of the repeated responses are determined. In this way, repeated responses can be better screened out, avoiding mistaking the subjective operations performed by users for repeated operations, thereby improving the effectiveness of the analysis results.

[0110] Example 5

[0111] On the basis of Example 1, the user satisfaction analysis method based on user event training further includes:

[0112] Identifying in each of the instant response information the first response information point corresponding to the first response destination;

[0113] Acquire, in the instant response information, a first information time difference between the first response information point and the corresponding instant response information end point and a second information time difference between the first response information point and the corresponding instant response information start point;

[0114] When the time difference of the first information is greater than the time difference of the second information, the corresponding immediate response information is regarded as delayed response information;

[0115] When the second information time difference is greater than the preset time difference, the corresponding immediate response information is regarded as delayed response information.

[0116] In this example, the preset time difference is 3 seconds.

[0117] The working principle and beneficial effects of the above technical solution: In order to further ensure the effectiveness of the classification of immediate response information and delayed response information, the immediate response information is refined and the delayed response information mixed therein is separated.

[0118] Example 6

[0119] On the basis of Example 1, the user satisfaction analysis method based on user event training, step 4 includes:

[0120] Step 41: determining the number of activity repetitions of the repeated activity event contained in the activity log according to the number of response repetitions corresponding to the activity response, establishing an event statistics table based on the number of activity repetitions, identifying the repeated activity events contained in the activity log, and filling the repeated activity events in the event statistics table according to the activity time corresponding to each repeated activity event;

[0121] Step 42: Analyze the activity time distribution information corresponding to each repeated activity event in the event statistics table, classify and annotate the event in the event statistics table, generate the annotated event table of the user, input the annotated event table into a preset neural network for deep training, obtain the valid activities and invalid activities contained in the distribution information of each activity event, and mark the effects in the corresponding repeated activity events respectively;

[0122] Step 43: Count the invalid information corresponding to each of the repeated activity events, determine the potential defects of the Web interface according to the invalid range corresponding to each of the invalid information, trace each of the potential defects respectively, obtain the historical defect remediation information of the Web interface, generate the user satisfaction report on the Web interface according to the potential defects and the corresponding historical defect remediation information, and display it.

[0123] In this example, valid activities refer to repeated activity events to which the Web system has made a correct response, and invalid activities refer to repeated activity events to which the Web system has not made a correct response, as well as activity events to which the Web system has made multiple repeated correct responses in a short period of time;

[0124] In this example, potential defects refer to defects that may occur in the web interface.

[0125] The working principle and beneficial effects of the above technical solution are as follows: in order to avoid statistical omissions, an event statistics table is established according to the number of activity repetitions of repeated events, and then the repeated activity events identified in the activity log are filled in the table, which can avoid event disorder and count the number of events, and then analyze the activity event distribution information of the repeated activity events, classify and label them in the event statistics table, and then use the preset neural network for deep training to determine the valid activities and invalid activities, and then count the invalid information in the repeated activity events, so as to determine the potential defects in the Web interface, and find the corresponding historical defect remediation information by tracing the defects, and finally generate a satisfaction report, which not only includes the user's satisfaction with each event, but also includes the remediation plan for the potential defects and the satisfaction that can be made up after the remediation. In this way, a complete and efficient report can be given to the management personnel, and the user's satisfaction can be continuously improved through remediation.

[0126] Example 7

[0127] Based on Example 6, the user satisfaction analysis method based on user event training further includes:

[0128] Acquire the suggestion information submitted by the user, and extract the user's usage defects from the suggestion information;

[0129] A defect list is generated and displayed according to the usage defects and potential defects of the Web interface.

[0130] The working principle and beneficial effects of the above technical solution are as follows: when a user submits a suggestion, the usage defects in the web interface are determined based on the submitted suggestion, and a list is generated for display to the management personnel for reference.

[0131] Example 8

[0132] Based on Example 1, the user satisfaction analysis method based on user event training includes:

[0133] Obtain satisfaction reports corresponding to a number of users, perform fusion analysis on the satisfaction reports, obtain the number of presentations corresponding to each of the potential defects, and generate a treatment priority corresponding to each of the potential defects.

[0134] The working principle and beneficial effects of the above technical solution are as follows: potential defects encountered by multiple users are processed first, which can improve the satisfaction of multiple users in the shortest possible time.

[0135] Example 9

[0136] This embodiment provides a user satisfaction analysis system based on user event training, such as Figure 2As shown, including:

[0137] An event collection module is used to collect activity events generated by users when operating on the Web interface, analyze the event results corresponding to each activity event, and generate a user activity log for the user;

[0138] A response analysis module, used to identify in the response log of the Web interface the immediate response information and the delayed response information generated by the Web interface in response to the user operation;

[0139] An abnormality identification module, used for inputting the immediate response information and the delayed response information into the user activity log for event matching, and obtaining a plurality of activity responses of the Web interface and a response repetition number corresponding to each activity response;

[0140] The defect determination module is used to search for repeated activity events in the activity log, use a neural network to perform annotation training on the repeated activity events, determine potential defects of the Web interface, generate a user satisfaction report on the Web interface and display it.

[0141] In this example, the activity event represents the event generated when the user performs activities in the Web network;

[0142] In this example, the event result indicates the result generated after the user operates the web interface;

[0143] In this example, the immediate response information indicates that the Web system responds to the user's operation in a short time, and the delayed response information indicates that the Web system does not respond to the user's operation in a short time, but responds after a long time.

[0144] In this example, both the immediate response information and the delayed response information may correspond to an activity event;

[0145] In this example, the neural network is initially manually annotated with data to record users with different proficiency levels using different functional modules of multiple systems (distinguished by primary routing) to generate event-time series. At the same time, the user can record whether he is satisfied with the page operation at any time based on the current time. When the user uses the system, the network predicts the current user satisfaction in real time. When the satisfaction exceeds (satisfied) or is lower than (threshold), a prompt pops up to prompt the user to choose whether he is satisfied with the current page. If the user chooses, the result of the current user operation sequence is recorded to generate an online data set, collect the online data set regularly, train the model twice, and iterate the above operations until the network converges to the optimal state.

[0146] The working principle and beneficial effects of the above technical solution are as follows: In order to objectively analyze the user's experience of using the Web system and improve the user's interest in using the system, the user's activity events on the Web interface are first collected, and the user's activity log is established by analyzing the event results corresponding to each activity event. Then, the immediate response information and delayed response information of the Web interface are identified in the response log of the Web interface, and the corresponding activity events are further matched for the immediate response information and the delayed response information, and repeated activity events are screened. Finally, the repeated activity events are labeled and trained using a neural network to find defects in the Web interface and generate a user satisfaction report on the Web interface. In this way, synchronous analysis can be performed when the user uses the Web interface, which solves the limitation of relying on questionnaires to analyze user satisfaction in the past. By summarizing the user's usage experience at all times, the defects of the system can be better determined so that adjustments can be made to improve the fluency of the system and improve user satisfaction.

[0147] Example 10

[0148] Based on Example 9, the user satisfaction analysis system based on user event training is characterized in that the defect determination module includes:

[0149] a preprocessing unit, configured to determine the number of activity repetitions of the repeated activity events contained in the activity log according to the number of response repetitions corresponding to the activity response, establish an event statistics table based on the number of activity repetitions, identify the repeated activity events contained in the activity log, and fill the repeated activity events in the event statistics table according to the activity time corresponding to each repeated activity event;

[0150] An event marking unit is used to analyze the activity time distribution information corresponding to each repeated activity event in the event statistics table, classify and mark in the event statistics table, generate a marked event table for the user, input the marked event table into a preset neural network for deep training, obtain the valid activities and invalid activities contained in the distribution information of each activity event, and mark the effects in the corresponding repeated activity events respectively;

[0151] A deep analysis unit is used to count the invalid information corresponding to each of the repeated activity events, determine the potential defects of the Web interface according to the invalid range corresponding to each of the invalid information, trace the source of each of the potential defects respectively, obtain the historical defect remediation information of the Web interface, generate the user satisfaction report on the Web interface according to the potential defects and the corresponding historical defect remediation information, and display it.

[0152] In this example, valid activities refer to repeated activity events to which the Web system has made a correct response, and invalid activities refer to repeated activity events to which the Web system has not made a correct response, as well as activity events to which the Web system has made multiple repeated correct responses in a short period of time;

[0153] In this example, potential defects refer to defects that may occur in the web interface.

[0154] The working principle and beneficial effects of the above technical solution are as follows: in order to avoid statistical omissions, an event statistics table is established according to the number of activity repetitions of repeated events, and then the repeated activity events identified in the activity log are filled in the table, which can avoid event disorder and count the number of events, and then analyze the activity event distribution information of the repeated activity events, classify and label them in the event statistics table, and then use the preset neural network for deep training to determine the valid activities and invalid activities, and then count the invalid information in the repeated activity events, so as to determine the potential defects in the Web interface, and find the corresponding historical defect remediation information by tracing the defects, and finally generate a satisfaction report, which not only includes the user's satisfaction with each event, but also includes the remediation plan for the potential defects and the satisfaction that can be made up after the remediation. In this way, a complete and efficient report can be given to the management personnel, and the user's satisfaction can be continuously improved through remediation.

[0155] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A user satisfaction analysis method based on user event training, characterized in that: include: Step 1: Collect activity events generated by users when operating on the Web interface, analyze the event results corresponding to each activity event, and generate a user activity log of the user; Step 2: Identify in the response log of the Web interface the immediate response information and the delayed response information generated by the Web interface in response to the user operation; Step 3: Input the immediate response information and the delayed response information into the user activity log for event matching, and obtain a number of activity responses of the Web interface and the number of response repetitions corresponding to each activity response; Step 4: Search for repeated activity events in the activity log, use a neural network to perform annotation training on the repeated activity events, determine potential defects of the Web interface, generate a user satisfaction report on the Web interface, and display it.

2. A user satisfaction analysis method based on user event training as claimed in claim 1, characterized in that: The step 1 comprises: Step 11: Collecting activity data generated when the user performs operation activities on the Web interface, establishing the user's access track according to the activity data, identifying the encryption level corresponding to each track point in the access track, and encrypting the corresponding data segment in the activity data according to the encryption level to obtain a plurality of encrypted data segments; Step 12: Establishing the activity event corresponding to the user according to the encrypted data segment, identifying the event start point and event end point corresponding to each activity event on the Web interface, arranging the activity events according to the event start point and event end point, and generating the event chain of the user; Step 13: Analyze the event logical relationship between different activity events in the event chain, determine the event result corresponding to each activity event in combination with the event start point and event end point corresponding to each activity event, and obtain the operation content and operation time corresponding to each event result to generate the user activity log of the user.

3. The user satisfaction analysis method based on user event training as claimed in claim 1, characterized in that: The step 2 comprises: Step 21: determining the operation time period of the user on the Web interface according to the user activity log, obtaining response data of the Web interface within the operation time period, generating a response log of the Web interface, and identifying the response information corresponding to each event result in the response log; Step 22: performing a first classification on the response information according to the operation action corresponding to each of the response information to generate a first response classification group, and performing a second classification on the response information according to the operation element corresponding to each of the response information to generate a second response classification group; Step 23: performing action training on each of the first response classification groups to obtain the corresponding operation pause preference when the user performs the corresponding operation action, and performing element recognition on each of the second response classification groups to obtain the basic response duration corresponding to each execution element; Step 24: respectively obtain the operation pause preference and basic response time corresponding to each of the response information, generate an immediate response range corresponding to the response information, analyze in the response log whether the actual response time of each of the response information is within the corresponding immediate response range, and divide the response information into immediate response information and delayed response information according to the analysis result.

4. A user satisfaction analysis method based on user event training as claimed in claim 1, characterized in that: The step 3 comprises: Step 31: extracting a first response key corresponding to each of the immediate response messages, and extracting a second response key corresponding to each of the delayed response messages, using the first response key to identify a first keyword of the response in the user activity log, and using the second response key to identify a second keyword of the response in the user activity log; Step 32: Determine a first activity event corresponding to each instant response message according to a matching relationship between the first response key and the first keyword, determine a second activity event corresponding to each instant response message according to a matching relationship between the second response key and the second keyword, establish a first response purpose according to the first activity event, and establish a second response purpose according to the second activity event; Step 33: identifying a first response information point corresponding to the first response destination in the immediate response information, treating the first response information point as the immediate response information end point of the immediate response information, generating an immediate active response of the Web interface, identifying a second response information point corresponding to the second response destination in the delayed response information, intercepting the information segment before the second response information point in the delayed response information, and generating a delayed active response of the Web interface; Step 34: The immediate activity response and the delayed activity response are regarded as activity responses, response attributes corresponding to each of the activity responses are obtained respectively, activity responses with consistent response attributes are regarded as repeated responses, and the number of response repetitions corresponding to each of the repeated responses is counted respectively.

5. The user satisfaction analysis method based on user event training according to claim 1, characterized in that: Also includes: Identifying in each of the instant response information the first response information point corresponding to the first response destination; Acquire, in the instant response information, a first information time difference between the first response information point and the corresponding instant response information end point and a second information time difference between the first response information point and the corresponding instant response information start point; When the time difference of the first information is greater than the time difference of the second information, the corresponding immediate response information is regarded as delayed response information; When the second information time difference is greater than the preset time difference, the corresponding immediate response information is regarded as delayed response information.

6. A user satisfaction analysis method based on user event training as claimed in claim 1, characterized in that: The step 4 comprises: Step 41: determining the number of activity repetitions of the repeated activity event contained in the activity log according to the number of response repetitions corresponding to the activity response, establishing an event statistics table based on the number of activity repetitions, identifying the repeated activity events contained in the activity log, and filling the repeated activity events in the event statistics table according to the activity time corresponding to each repeated activity event; Step 42: Analyze the activity time distribution information corresponding to each repeated activity event in the event statistics table, classify and annotate the event in the event statistics table, generate the annotated event table of the user, input the annotated event table into a preset neural network for deep training, obtain the valid activities and invalid activities contained in the distribution information of each activity event, and mark the effects in the corresponding repeated activity events respectively; Step 43: Count the invalid information corresponding to each of the repeated activity events, determine the potential defects of the Web interface according to the invalid range corresponding to each of the invalid information, trace each of the potential defects respectively, obtain the historical defect remediation information of the Web interface, generate the user satisfaction report on the Web interface according to the potential defects and the corresponding historical defect remediation information, and display it.

7. A user satisfaction analysis method based on user event training as claimed in claim 6, characterized in that: Also includes: Acquire the suggestion information submitted by the user, and extract the user's usage defects from the suggestion information; A defect list is generated and displayed according to the usage defects and potential defects of the Web interface.

8. The user satisfaction analysis method based on user event training as described in claim 1 is characterized in that: include: Obtain satisfaction reports corresponding to a number of users, perform fusion analysis on the satisfaction reports, obtain the number of presentations corresponding to each of the potential defects, and generate a treatment priority corresponding to each of the potential defects.

9. A user satisfaction analysis system based on user event training, characterized in that: include: An event collection module is used to collect activity events generated by users when operating on the Web interface, analyze the event results corresponding to each activity event, and generate a user activity log for the user; A response analysis module, used to identify in the response log of the Web interface the immediate response information and the delayed response information generated by the Web interface in response to the user operation; An abnormality identification module, used for inputting the immediate response information and the delayed response information into the user activity log for event matching, and obtaining a plurality of activity responses of the Web interface and a response repetition number corresponding to each activity response; The defect determination module is used to search for repeated activity events in the activity log, use a neural network to perform annotation training on the repeated activity events, determine potential defects of the Web interface, generate a user satisfaction report on the Web interface and display it.

10. A user satisfaction analysis system based on user event training as claimed in claim 9, characterized in that: The defect determination module comprises: a preprocessing unit, configured to determine the number of activity repetitions of the repeated activity events contained in the activity log according to the number of response repetitions corresponding to the activity response, establish an event statistics table based on the number of activity repetitions, identify the repeated activity events contained in the activity log, and fill the repeated activity events in the event statistics table according to the activity time corresponding to each repeated activity event; An event marking unit is used to analyze the activity time distribution information corresponding to each repeated activity event in the event statistics table, classify and mark in the event statistics table, generate a marked event table for the user, input the marked event table into a preset neural network for deep training, obtain the valid activities and invalid activities contained in the distribution information of each activity event, and mark the effects in the corresponding repeated activity events respectively; A deep analysis unit is used to count the invalid information corresponding to each of the repeated activity events, determine the potential defects of the Web interface according to the invalid range corresponding to each of the invalid information, trace the source of each of the potential defects respectively, obtain the historical defect remediation information of the Web interface, generate the user satisfaction report on the Web interface according to the potential defects and the corresponding historical defect remediation information, and display it.

Images