WO2020233309A1 - 软件产品测评处理方法、装置、计算机设备及存储介质 - Google Patents

软件产品测评处理方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2020233309A1
WO2020233309A1 PCT/CN2020/085579 CN2020085579W WO2020233309A1 WO 2020233309 A1 WO2020233309 A1 WO 2020233309A1 CN 2020085579 W CN2020085579 W CN 2020085579W WO 2020233309 A1 WO2020233309 A1 WO 2020233309A1
Authority
WO
WIPO (PCT)
Prior art keywords
comment
target
data
keyword
question
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/085579
Other languages
English (en)
French (fr)
Inventor
赵昊
赵晔菲
张佩茜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
OneConnect Smart Technology Co Ltd
Original Assignee
OneConnect Smart Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by OneConnect Smart Technology Co Ltd filed Critical OneConnect Smart Technology Co Ltd
Publication of WO2020233309A1 publication Critical patent/WO2020233309A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/205—Parsing
    • G06F40/216—Parsing using statistical methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0623—Electronic shopping [e-shopping] by investigating goods or services

Definitions

  • This application relates to the technical field of financial tools, and in particular to a software product evaluation processing method, device, computer equipment and storage medium.
  • this software product evaluation process has the following deficiencies: First, the cost is relatively high and the efficiency is low, and the product development organization needs to be equipped with corresponding personnel for evaluation problem design and subsequent evaluation analysis, and its labor cost And the time cost is high, and the efficiency is low; the second is that the real-time performance is not strong. Generally, only the comment data of the commentator during the evaluation period is collected, and it is impossible to collect the commenters’ thoughts or problems in the process of using the software product in real time. Product optimization; the third is that the analysis results have limitations. Because the review data is generally based on self-set evaluation questions, the analysis results are centered around the evaluation questions and cannot reflect the reviewers’ thoughts in the process of using software products in a timely manner.
  • the embodiments of the application provide a software product evaluation processing method, device, computer equipment, and storage medium to solve the problems of high cost, low efficiency, low real-time performance and limited analysis results in the current software product evaluation process problem.
  • a software product evaluation processing method including:
  • target comment data does not carry the task identifier, use a keyword extraction algorithm to perform keyword extraction on the target comment data to obtain the target keyword corresponding to the target comment data;
  • the question keyword is determined as a high frequency keyword, and the corresponding high frequency question is determined based on the target comment data containing the high frequency keyword;
  • a software product evaluation and processing device including:
  • the target comment data obtaining module is used to create a target comment group corresponding to any specific software product installed on the product evaluation platform based on the instant messaging tool, and obtain target comment data and corresponding target comment data published by any target commentator in the target comment group To determine whether the target comment data carries a task identifier;
  • the target keyword acquisition module is configured to, if the target comment data does not carry the task identifier, use a keyword extraction algorithm to perform keyword extraction on the target comment data to obtain the target keyword corresponding to the target comment data;
  • Question keyword acquisition module configured to determine the question keyword corresponding to the target comment data according to the target keyword corresponding to the target comment data and the comment time;
  • An appearance frequency obtaining module configured to perform frequency statistics on the question keywords in the comment statistical period corresponding to the current time of the system, and obtain the appearance frequency corresponding to the question keywords;
  • a high-frequency question determination module configured to determine the question keyword as a high-frequency keyword if the frequency of occurrence is greater than a preset frequency threshold, and determine the corresponding high-frequency keyword based on the target comment data containing the high-frequency keyword problem;
  • the target priority acquiring module is used to acquire the frequency of occurrence of the problem and the degree of impact of the problem corresponding to the high-frequency problem, perform priority analysis based on the frequency of occurrence of the problem and the degree of impact of the problem, and obtain the target corresponding to the high-frequency problem priority;
  • the response processing module is configured to query the comment response mechanism information table based on the target priority, obtain the corresponding comment response mechanism, and respond to the high-frequency question based on the comment response mechanism.
  • a computer device includes a memory, a processor, and a computer program that is stored in the memory and can run on the processor, and the processor implements the above-mentioned software product evaluation processing method when the computer program is executed.
  • a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned software product evaluation processing method is realized.
  • the above software product evaluation processing method, device, computer equipment and storage medium create a target review group corresponding to a specific software product through instant messaging tools, and obtain target review data published by any target reviewer in the target review group, thereby ensuring Real-time processing of target comment data.
  • FIG. 1 is a schematic diagram of an application environment of a software product evaluation processing method in an embodiment of the present application
  • Figure 2 is a flowchart of a software product evaluation processing method in an embodiment of the present application
  • FIG. 5 is another flowchart of the software product evaluation processing method in an embodiment of the present application.
  • FIG. 6 is another flowchart of the software product evaluation processing method in an embodiment of the present application.
  • FIG. 7 is a schematic diagram of a software product evaluation processing device in an embodiment of the present application.
  • Fig. 8 is a schematic diagram of a computer device in an embodiment of the present application.
  • the software product evaluation processing method provided by the embodiment of the application can be applied to the application environment shown in FIG. 1.
  • the software product evaluation processing method is applied in a product evaluation platform, and the product evaluation platform is a comprehensive platform for realizing software product evaluation.
  • the product evaluation platform includes the client and server as shown in Figure 1.
  • the client and server communicate through the network to realize the research and analysis of software products, which not only ensures the pertinence of the reviewers’ comments and the flexibility of the review time, In order to optimize software products based on the final target comment results obtained, and effectively reduce the software product optimization cycle.
  • the client is also called the client, which refers to the program that corresponds to the server and provides local services to the client.
  • the client can be installed on, but not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices.
  • the server can be implemented as an independent server or a server cluster composed of multiple servers.
  • a software product evaluation processing method is provided, which is applied on a product evaluation platform.
  • the method is applied on the server shown in Figure 1, and specifically includes the following steps:
  • S201 Create a target comment group corresponding to any specific software product installed on the product evaluation platform based on the instant messaging tool, obtain target comment data and corresponding comment time published by any target commentator in the target comment group, and determine the target comment data Whether to carry the task ID.
  • the target comment group is an instant communication group created based on instant messaging tools and used to comment on specific software products.
  • the target comment group is an instant messaging group composed of system administrators and target commentators for commenting on specific software products. Communication group.
  • the specific software product is installed on the product evaluation platform, so that the product evaluation platform can evaluate the specific software product, ensuring the executable performance of the comprehensive evaluation of the software product, without the need for the product development organization to allocate corresponding personnel for evaluation Processing, saving labor costs.
  • the system administrator is a group member used to create and manage a target comment group, mainly used for posting comment tasks or collecting comment data, and is generally a staff member of a product development organization.
  • the target commentator is a group member in the target comment group.
  • the target commentator is a group member who meets the preset commentator conditions, and is mainly used to publish comment data, so that the target commentator who publishes the target comment data has The fixedness saves the time for determining the target commentator and helps ensure the pertinence of the target comment data.
  • the preset commentator conditions may specifically include conditions such as the degree of active participation in the activity and whether there is intention.
  • Each target comment data corresponds to a comment time, and the comment time is the time when the target commentator publishes the target comment data on the target comment group.
  • the target comment data is data published by the target commentator on the target comment group, which helps to ensure the timeliness of comment data analysis.
  • the server obtains target comment data published by any target commentator in the target comment group, specifically including the following two situations:
  • the first method specifically includes the following steps: 1-1)
  • the server obtains the task to be commented published by the system administrator through the comment terminal, and displays the task to be commented on the target comment group of the client.
  • the task to be commented includes task identification, task content and Task comment period.
  • the server obtains the target comment data corresponding to the task to be commented in a specific comment data format published by any target commentator in the target comment group through the comment terminal, and the target comment data corresponds to a comment time.
  • the target comment data in the specific comment data format carries the task identifier.
  • the comment terminal is a terminal used for system administrators or target commentators to post information on the target comment group, and may be a terminal installed with an instant messaging client.
  • the task to be commented is a task that needs to be commented created based on a special issue led by the product development organization, that is, the task to be commented is a task issued for a research question pre-designed by the product development organization.
  • the task ID is used to distinguish a certain comment task from other comment tasks. It can be set according to the preset task numbering rule.
  • the task numbering rule can be determined by the task label and the sequence label, for example, in RW001 and RW002, RW is a task label, and 001 and 002 are sequence labels.
  • each task identifier corresponds to a task to be reviewed.
  • the content of the task is used to explain the subject of the comment that the task to be commented on, such as the experience of using a certain function, pain points or thoughts on new functions, etc.
  • the task comment period is used to limit the comment period of the task to be commented, so as to prompt the target commentator to comment in time to ensure the efficiency of software product optimization.
  • the system administrator can post the task to be commented in the target comment group, so that all target commentators in the target comment group can comment on the task to be commented to obtain the target comment in a specific comment data format. data.
  • a specific comment data format may be configured in advance, so that the target commentator uses the specific comment data format to post corresponding target comment data.
  • the specific comment data format may be: comment identification (such as #)-task identification (such as RW001 or RW002)-comment content data. Understandably, target commentators use this specific comment data format to publish target comment data, which can help to uniformly analyze and process target comment data corresponding to the same comment identifier, improve the efficiency of software product research, and shorten software product optimization. cycle.
  • the target commenter can input the corresponding comment identifier (such as #) through the comment terminal ,
  • the review terminal displays a list of tasks to be reviewed for all tasks to be reviewed that are currently within the task review period, so that the target reviewer can select the tasks to be reviewed based on the list of tasks to be reviewed.
  • the list of tasks to be reviewed Each task identification and corresponding task content are displayed on the screen.
  • the target commentator After the target commentator selects any task to be commented in the task list to be commented, and then enters the comment content data through the comment terminal, the "comment identifier (such as #)-task identifier (such as RW001 or RW002)-comment content data" can be obtained.
  • the target comment data in the specific comment data format is used to ensure the publication efficiency of the target comment data in the specific comment data format.
  • the second method specifically includes the following steps: 2-1)
  • the server obtains the target comment data that is randomly published by any target commentator in the target comment group based on the comment terminal without adopting the specific comment data format. That is, in this case, the target comment data that does not adopt the specific comment data format does not carry the task identifier.
  • Target comment data when the target reviewer has a new idea in the process of using the specific software product or thinks that there is a certain problem or pain point, which may be useful for the improvement of the software product function, he can post through the review terminal that the specific review data format is not adopted.
  • Target comment data can be a question-type feedback on the existing functions of the specific software product, or it can be a new demand-type feedback independently proposed, that is, the target comment data is not a comment on a task to be commented by the system administrator.
  • the data Understandably, by analyzing the target comment data randomly published by any target commentator in the target comment group, the comment time can be made flexible, which is helpful to summarize and analyze the new function plan or new software product optimization and improvement. Ideas, new functional planning and new ideas are optimized and perfected to optimize software products, avoiding problems with directly optimized software products and requiring repeated revisions, thereby shortening the optimization cycle of software products.
  • the server can use a string matching algorithm or a regular expression matching algorithm to process the target comment data to determine whether the target comment data carries a task identifier .
  • a string matching algorithm or a regular expression matching algorithm to process the target comment data to determine whether the target comment data carries a task identifier .
  • the server can use a string matching algorithm or a regular expression matching algorithm to process the target comment data to determine whether the target comment data carries a task identifier .
  • the server can use a string matching algorithm or a regular expression matching algorithm to process the target comment data to determine whether the target comment data carries a task identifier .
  • the target keyword is a keyword that can be used to reflect the target comment data after keyword extraction is performed on the target comment data.
  • the keyword extraction algorithm is an algorithm used to extract keywords in the text. It can use but is not limited to the TextRank algorithm, or can segment the target comment data and remove stop words to determine its corresponding target keyword.
  • the target comment data does not carry the task identifier, it means that the target comment data is not based on the comments published by the system administrator on the task to be commented in advance, but the new ideas generated by the target commentator in the daily use of the software product , Pain points or problems and other content feedback comments.
  • the target comment data published by the target comment data may be question-type feedback or demand-type feedback.
  • the problem feedback refers to the feedback on the existing functions of the software product, the problems that occur in the product and the operation process.
  • Demand feedback refers to the feedback of new features that the target reviewer expects the software product to add.
  • the server uses a keyword extraction algorithm to extract keywords from the target comment data published by the target commentator, so as to extract target keywords that reflect the key content of the target comment data , In order to carry out the statistical analysis and processing of the data subsequently, improve the processing efficiency.
  • S203 Determine the question keyword corresponding to the target comment data according to the target keyword and the comment time corresponding to the target comment data.
  • the target comment data does not carry the task identifier, it means that the target comment data published by the target commentator is not a comment on the task to be commented in advance released by the system administrator. At this time, the comment question targeted by the target comment data cannot be directly determined.
  • the specific content of the target comment data needs to be analyzed to determine the comment question targeted by the target comment data.
  • the question keyword corresponding to the target comment data specifically refers to the keyword corresponding to the comment question targeted by the target comment data. For example, in the target comment data of "The login interface of this APP is easy to freeze and the experience is not good", the question keyword can be "login interface", that is, the question keyword is mainly for a certain functional interface of a specific software product. To determine that the subject of the comment is a problem for this functional interface.
  • step S203 specifically includes the following steps:
  • the part of speech tagging tool is a tool for tagging the part of speech of any target keyword, including but not limited to the PosTagger tool.
  • S2032 Determine the target keyword whose part of speech is the preset part of speech as the keyword to be analyzed.
  • the preset part of speech is the part of speech that needs to be analyzed in advance, such as nouns, verbs and adjectives.
  • the question keyword database is a preset database for storing preset keywords.
  • the preset keywords are preset keywords related to the comment question.
  • the matching algorithm is a preset algorithm for matching whether the two are the same or similar, including but not limited to a regular expression matching algorithm. For example, if a regular expression matching algorithm is used to match a preset keyword that is the same as the keyword to be analyzed, the obtained matching result is a successful match, otherwise, the obtained matching result is a failed match.
  • the target keywords extracted by matching the target comment data include the question keywords "login interface” and "stuck”.
  • the question keywords can be directly determined; when the comment question is not clearly described in the target comment data, the last history before the comment time of the target comment data
  • the problem keywords of the comment data are determined as the problem keywords of the target comment data.
  • S204 Perform frequency statistics on the question keywords in the comment statistical period corresponding to the current time of the system, and obtain the occurrence frequency corresponding to the question keywords.
  • the current system time refers to the system time of the server.
  • the comment statistical period is a preset period for statistical analysis of the question keywords corresponding to the target comment data.
  • the comment statistical period can be one week or other times.
  • the comment statistical period corresponding to the current time of the system refers to the time interval corresponding to the comment statistical period from the current time of the system, that is, the current time of the system is the statistical cut-off time to ensure the real-time statistics of the occurrence frequency of problem keywords. For example, if the current time of the system is February 10, and the comment statistical period is 3 days, the comment statistical period corresponding to the current time of the system is February 8, February 9, and February 10.
  • the question keywords extracted from the target comment data also correspond to the comment time.
  • the frequency statistics of the question keywords in the comment statistical period corresponding to the current time of the system are performed to obtain the occurrence frequency corresponding to the question keywords, which specifically includes: the server first obtains the comments corresponding to the comment time in the current system time For all target comment data in the statistical period, the number of times the target comment data contains the question keyword is counted to obtain the frequency of occurrence of the question keyword.
  • the preset frequency threshold is a preset threshold for evaluating whether the number of occurrences identified as high frequency is reached. Specifically, the server compares the occurrence frequency corresponding to the problem keyword with a preset frequency threshold value set in advance, and if the occurrence frequency corresponding to the problem keyword is greater than the preset frequency threshold value, it is determined that the problem keyword is a high-frequency keyword.
  • the comment question extracted from the target comment data corresponding to the high frequency keyword is determined as a high frequency question.
  • by comparing the occurrence frequency of the question keyword in the comment statistical period corresponding to the current time of the system with the preset frequency threshold it is determined whether the question keyword is a high-frequency keyword, based on the fact that the high-frequency keyword is included.
  • To determine the high-frequency problem corresponding to the target review data to ensure the timeliness of the determination of the high-frequency problem (both data in the review statistical period corresponding to the current time of the system), which helps to improve the optimization efficiency of software products.
  • determining the corresponding high-frequency question based on the target comment data containing the high-frequency keyword specifically includes: if there is only one target comment data containing the high-frequency keyword, directly according to the target comment data contained in the target comment data
  • S206 Count the frequency of occurrence of the problem and the degree of impact of the problem corresponding to the high-frequency problem, and perform priority analysis according to the frequency of occurrence of the problem and the degree of problem impact, and obtain the target priority corresponding to the high-frequency problem.
  • the problem occurrence frequency is an analysis result of the occurrence frequency determined according to the response of the high-frequency problem in the target comment group. Specifically, after the server determines that a certain comment question is a high-frequency question, it publishes the corresponding information response task on the target comment group, and obtains the question response information of all target commentators in the target comment group for the information response task, based on the question Respond to the statistics of the probability of the problem, and compare the probability of the problem with the preset probability threshold. If the probability of the problem is greater than the preset probability threshold, obtain the probability of the problem with a greater probability of occurrence; if the probability of the problem is not greater than If the probability threshold is preset, the problem occurrence probability result with a small problem occurrence probability is obtained.
  • the information response task is a task issued by the system administrator to collect whether the target commentator encounters a certain high-frequency problem.
  • Question response information is the information that responds to the information response task in the target comment.
  • the preset probability threshold is a preset probability-related threshold.
  • target commentator A puts forward a target comment data and considers it to be a high-frequency question.
  • the system administrator publishes an information response task based on the high-frequency question, if the question response is obtained
  • the information is that there are 100 target commentators who responded that they have encountered corresponding problems.
  • the probability of occurrence of the problem is 10%. Compare the probability of occurrence of the problem with a preset probability threshold (such as 5%) , It can be determined that the corresponding problem occurrence frequency is the higher probability of occurrence.
  • the degree of problem impact is an index used to reflect the degree of impact of high-frequency problems on user experience.
  • the degree of impact of the problem can be determined according to the severity of the problem.
  • the severity level of the problem adopts Wilson (1990)'s five-level standard as follows: level 5-insignificant errors; level 4-the problem is small but makes users anxious; level 3-moderate, time-consuming but No data will be lost; level 2-serious problems that cause data loss; level 1-catastrophic errors that cause data loss or damage to software and hardware.
  • the severity level of the problem can use another five-level standard as follows: 0-not considered as a usability problem; 1-this is just a usability problem of decorative facades: no special treatment is required, unless this item has additional Time; 2-Minor availability problem: the priority of solving this problem is low; 3-Major availability problem: solving this problem is very important, the priority is high; 4- availability disaster ( Catastrophe): It is necessary and urgent to solve this problem (Imperative), and it must be solved before the software product is released.
  • the degree of impact of the problem matches the severity level of the problem, for example, including five levels of problem impact.
  • the degree of impact of the problem can be calculated and determined based on the severity of the problem.
  • the severity of the five-level is divided into two levels of problem impact, such as the five-level standard of Wilson (1990), level 5 and level 4
  • the standard problem is determined as the problem impact degree with less impact degree
  • the problem with the level 1, 2 and 3 standards is determined as the problem impact degree with greater problem impact degree.
  • performing priority analysis on high-frequency issues to obtain target priorities corresponding to high-frequency issues specifically includes: performing priority analysis on high-frequency issues to obtain the occurrence frequency and degree of impact of the issues corresponding to the high-frequency issues; Query the priority information comparison table based on the frequency of the problem and the degree of the problem's impact to obtain the target priority corresponding to the high-frequency problem.
  • the priority information comparison table is a comparison table of priority information used to reflect the combination of different problem occurrence frequency and problem impact degree (as shown in Table 1 below).
  • the priority analysis of high-frequency issues refers to the comprehensive analysis of the two evaluation indicators, namely, the frequency of occurrence of high-frequency issues and the degree of impact of the issues, based on the pre-set evaluation index judgment criteria, to determine their corresponding Goal prioritization process.
  • the target priority refers to the priority determined after comprehensive analysis based on the frequency of the problem and the degree of impact of the problem. The target priority feedbacks the urgency of high-frequency problems to a certain extent.
  • the comment response mechanism information table is an information table used to reflect the response mechanism of the high-frequency question corresponding to each priority.
  • the response mechanism is a mechanism that specifies the response time and response process for a certain high-frequency problem.
  • Response time refers to the time to respond to high-frequency problems.
  • the response process can specifically include the process of responding to a high-frequency problem configuration and the corresponding processing personnel, so that the response to high-frequency problems is more reasonable and efficient, to ensure the efficiency of high-frequency problem processing, and to improve the software product Optimize efficiency and shorten the optimization cycle of software products.
  • an instant messaging tool is used to create a target comment group corresponding to any specific software product installed on the product evaluation platform, and obtain any target commentator in the target comment group.
  • Target comment data to ensure the real-time processing of target comment data.
  • the target comment data does not carry the task identifier, it means that the target comment data is not a comment published for the task to be commented, which makes the comment content of the target comment data more flexible and avoids limitations.
  • a keyword extraction algorithm is used to extract keywords from the target comment data to obtain the target keywords to ensure the efficiency of subsequent data processing.
  • the corresponding question keyword is determined, thereby determining the comment question corresponding to each target comment data, and ensuring the objectivity of the follow-up analysis to determine the high-frequency question.
  • determine whether the question keyword is a high-frequency keyword so as to determine the high-frequency problem to ensure that the high-frequency problem is determined
  • the timeliness of this helps to improve the optimization efficiency of software products.
  • the software product evaluation processing method further includes:
  • the target comment data carries the task identifier
  • the server needs to obtain the comment time corresponding to the target comment data, and query the list of tasks to be commented based on the task identifier, obtain the task comment period corresponding to the task identifier, and determine the target comment data Whether the comment time of is within the comment period of the task, in order to analyze whether the target comment data is valid, so as to ensure the timeliness of the task to be commented.
  • the task to be commented list is a data table used to record task information of all tasks to be commented on, and is specifically used to store task information such as task identification, task content, and task comment period corresponding to each task to be commented on.
  • the comment time of the target comment data is within the task comment period corresponding to the task identifier, it means that the target commentator made a comment during the task comment period of the task to be commented corresponding to the task identifier, and is for the task to be reviewed.
  • the target comment data can be used as a basis for subsequent data analysis, that is, effective comment data.
  • the effective review data refers to the review data that can be used as a basis for data analysis during the review period of the corresponding task.
  • the server's analysis of the effective comment data refers to an analysis process used to analyze the comment question targeted by each effective comment data and what tendency the target commentator has towards this comment question.
  • the server may use a natural language processing tool to analyze the effective comment data to obtain comment questions corresponding to the effective comment data and comment tendency results corresponding to the comment questions.
  • the comment question refers to the question targeted by the valid comment data, and the comment question can match the task content in the task to be commented. For example, if the target comment question is clearly mentioned in the valid comment data, the comment question corresponding to the valid comment data can be directly determined; if the target comment question is not clearly mentioned in the valid comment data (such as omitting the question and quoting it directly) Effective comment data published in the form of the task to be commented), the corresponding comment question can be determined according to the content of the task in the task to be commented.
  • the comment tendency result is the result of extracting a certain tendency of the target commentator from the effective comment data published by the target commentator. For example, if the effective review data is data for reviewing a review question of whether the data upload function of a specific software product is useful or not, then the review tendency result includes two different tendencies: good use and poor use. If the valid comment data is for the three versions of the data upload function of P1, P2, and P3, which is the more useful comment question to evaluate, then the comment tendency result includes the comment tendency result of P1, P2 or P3 which is more useful.
  • the server counts the number of comments corresponding to each comment tendency result corresponding to the same comment question in the effective comment data corresponding to the task to be commented corresponding to the same task identifier in the target comment group, and the number of comments can be understood as the target comment group The number of target commentators who support the comment preference result. Then, the server determines the comment tendency result with the largest number of comments as the target comment result corresponding to the comment question.
  • the valid comment data is for the three versions of the data upload function of P1, P2, and P3, evaluate which comment is more useful.
  • Different target commentators have different comment tendency results, and the server needs to respond to this comment question.
  • a unified summary of the results of each comment tendency to obtain the number of comments corresponding to each comment tendency result to determine the comment tendency result with the largest number of comments as the target comment result corresponding to the comment question, that is, the majority principle is adopted to ensure the target
  • the objectivity of the review results For example, among 1000 valid comment data, the number of comments that think the data upload function of the three versions of P1, P2 and P3 are more useful are 200, 300, and 500 respectively, then the data upload function of the P3 version is better to use.
  • the target comment result corresponding to the comment question are 1000 valid comment data, the number of comments that think the data upload function of the three versions of P1, P2 and P3 are more useful are 200, 300, and 500 respectively, then the data upload function of the P3 version is better to use.
  • the target comment data published by any target commentator in the target comment group is obtained, so that the commentator who publishes the target comment data is fixed, which helps to ensure the target comment data is targeted. Sex.
  • the target comment data carries the task identifier, if the comment time corresponding to the target comment data is within the task comment period corresponding to the task identifier, the target comment data is determined as valid comment data to ensure the comment timeliness of the valid comment data, which is helpful To improve the timeliness of comment data processing, thereby shortening the optimization cycle of software products.
  • the server can preset corresponding timing analysis tasks for realizing timing analysis of the comment behavior of each target commentator.
  • the timing analysis task is a task set in advance for executing the corresponding analysis process at a certain timing analysis time.
  • the timing analysis time is a preset time for analysis processing.
  • the historical comment data refers to the target comment data published by the target commentator before the current time of the system that can be used for data statistical analysis.
  • the user account of a target commentator refers to the unique account of each target commentator in the target comment group, which can be used to distinguish target comment data published by different target commentators.
  • the comment database is a database for storing target comment data published by all target commentators in the target comment database. Understandably, each target comment data stored in the comment database is stored in association with a user account, so that the comment behavior of each target commentator can be comprehensively analyzed.
  • the data statistical period is a preset period for statistical analysis of the target comment data of any target commentator.
  • the data statistical period corresponding to the timing analysis time refers to the time interval corresponding to the data statistical period before the timing analysis time. For example, if the timing analysis time is 12:00 every Monday, and the data statistical period is one week, the data statistical period corresponding to the timing analysis time refers to the week before the timing analysis time.
  • step S301 specifically includes: when the current time of the system is the timing analysis time, the server executes a preset timing analysis task, that is, first query the comment database based on the user account of each target commentator, and obtain the corresponding user account All target comment data and determine the comment time corresponding to each target comment data; then, determine the target comment data in the data statistical period corresponding to the timing analysis time as the corresponding user account that can be used for data statistical analysis Historical comment data.
  • a preset timing analysis task that is, first query the comment database based on the user account of each target commentator, and obtain the corresponding user account All target comment data and determine the comment time corresponding to each target comment data; then, determine the target comment data in the data statistical period corresponding to the timing analysis time as the corresponding user account that can be used for data statistical analysis Historical comment data.
  • S302 Query a list of comment items corresponding to the data statistical period based on the historical comment data, and obtain a comment participation ratio corresponding to the historical comment data.
  • the comment item list corresponding to the data statistical period is an information table used to record all comment items in the data statistical period.
  • the review item includes, but is not limited to, the item corresponding to the task to be reviewed independently determined by the system administrator, and also includes the item corresponding to the newly added review task created based on high-frequency questions.
  • the comment participation ratio is the ratio of the number of participating items used to feedback the target commentator's participation in the comment to the total number of items of all items in the entire review item list. For example, if there are 20 review items in a data statistical period, that is, the total number of items is 20, and the number of participating items in which a target commentator participates in the review is 16, then the obtained review participation rate is 80%. Understandably, the higher the comment participation ratio, the more actively the target commentator participates in the comment on the comment item, which can objectively reflect the active degree of the target commentator in the target comment group.
  • each target comment data has been analyzed in advance, and the comment tendency result corresponding to the historical comment data can be directly used as the historical tendency result. Accordingly, the target of the comment question pair can be directly obtained Comment on the result.
  • the historical comment data can be analyzed to determine the comment question corresponding to the historical comment data and the historical tendency result corresponding to the comment question. The process is as shown in step S204. To avoid repetition, I will not repeat them one by one here; and refer to step S205 to determine the target comment result corresponding to the comment question.
  • the available comment ratio refers to the ratio of the matching number of historical comment data whose historical tendency result matches the target comment result to the number of participating projects corresponding to all historical comment data. For example, if in a data statistical period, the number of participating items corresponding to all historical review data corresponding to a target reviewer is 16, among them, the historical tendency results of 12 review items are the same as the final target review results, that is, matching If the number is 12, the calculated comment available ratio is 75%. Understandably, the higher the available rate of the review, the greater the effect of the target reviewer's review suggestion on software product optimization, that is, the higher the usefulness of the review.
  • step S304 specifically includes the following steps: (1) The server queries the preset score comparison table based on the comment participation rate and the comment available rate corresponding to the historical comment data, and respectively determines the participation score P1 and the comment available rate corresponding to the comment participation rate.
  • the corresponding available score is P2.
  • the score comparison table is a preset comparison table for specifying the relationship between the participation ratio of each review and the corresponding participation score, and the correlation between the available ratio of the review and the corresponding available score.
  • P1 and P2 use scores in the same numerical range, so that both have comparable values, such as a value between 0-10, so that the calculated comment score is also between 0-10 Numerical value.
  • the review reward information includes, but is not limited to, cash red envelope rewards, member privilege rewards, and entity gift rewards, which can be independently determined according to different review scores P. Understandably, the larger the review score P, the richer the reward in the corresponding review reward information, the more helpful it is to attract the target commentator to actively publish the target comment data and mobilize the enthusiasm of the target commentator.
  • the determination of the review reward information is determined after comprehensive analysis and calculation based on the review participation proportion and the review available proportion corresponding to the historical review data, so as to make it objective.
  • the review data is queried through the user account of the target reviewer to determine the historical review data corresponding to the data statistical period, so that the review behavior analysis based on the historical review data is periodic , which is more conducive to prompting target commentators to actively publish target comment data.
  • the corresponding comment participation ratio and comment availability ratio are determined to objectively evaluate target commentators from the two dimensions of activity and comment usefulness. Based on the comment participation ratio and comment availability ratio corresponding to the historical comment data, the corresponding comment reward information is obtained to ensure the objectivity of the comment reward information, which is more helpful to attract target commentators to actively publish target comment data on the target comment group .
  • step S201 creating a target comment group corresponding to any specific software product installed on the product evaluation platform based on the instant messaging tool includes the following steps:
  • the comment group creation request is used to trigger a request to create a comment group for a specific software product on the product evaluation platform.
  • Product identification is used to uniquely distinguish between different software products.
  • the product attribute type is used to limit the attributes of the software product to be reviewed in the review group to be created this time.
  • the product attribute type includes two related types or subdivision function types. The system administrator can independently choose and determine the corresponding product attribute type according to actual needs.
  • Existing related types can be used to describe the corresponding types of software products targeted by this comment group creation request, such as financial management types, insurance types, or other types.
  • existing software products of the same type such as property insurance A and property insurance B
  • similar types such as property insurance A and life insurance C
  • select the existing related types as the products for the comment group creation request The attribute type.
  • the main reason is that when these existing software products have accumulated a considerable number of existing users, the reviewers to be recruited can be determined based on the user behavior data of the existing users, so that the determination of the reviewers to be recruited is more targeted.
  • the subdivision function type is used to describe the specific functions of the software product targeted by this comment group creation request.
  • the type is used as the product attribute type of the comment group creation request. Specifically, by disassembling the product functions of a specific software product to obtain the specific subdivision functions (such as login function, payment function, camera function or other functions) that the specific software product includes, and then looking for the corresponding subdivision function Of existing users in order to determine reviewers to be recruited based on the user behavior data of existing users.
  • S402 Create a target comment group corresponding to the product identification based on the instant messaging tool.
  • an instant messaging tool is an interface set on a server that can create an instant messaging group.
  • the instant messaging tool includes, but is not limited to, a WeChat interface, NetEase Yunxin, or other interfaces that can implement instant messaging functions.
  • the server creates a target comment group corresponding to the product identifier that can realize the instant messaging function based on a preset instant messaging tool, that is, the target comment group is a communication group that mainly reviews software products corresponding to the product identifier, so that the The purpose of creating the target comment group is pertinent.
  • S403 Query the user behavior database based on the product attribute type, and determine active users matching the product attribute type within a preset time period as commenters to be recruited.
  • the user behavior database is a database used to record user behavior data of existing users in related software products.
  • User behavior data is used to record the log-in time, log-in duration, and log-in functions of existing users in related software products.
  • Active users refer to users who have reached the active index based on user behavior data such as login time, login duration, and login functions of existing users.
  • the preset time period is the time set in advance to limit the user behavior data to be collected, so as to avoid identifying users who were active a long time ago (such as 1 year ago) and inactive for a period of time (such as 3 months) as Reviewers to be recruited can help ensure the enthusiasm of the target reviewers to participate in the review activities based on the target reviewers to be recruited.
  • Commentators to be recruited refer to existing users who are more likely to be recruited as final target commentators based on the active level of existing users participating in activities.
  • the server queries the corresponding user behavior database based on the product attribute type, and obtains the user behavior data of all existing users matching the product attribute type; and then filters out that the login time is within a preset time period corresponding to the current time of the system (such as The user behavior data in the past three months); finally, based on the filtered user behavior data, whether the existing user is an active user is analyzed to determine the active user as a commentator to be recruited.
  • whether an existing user is an active user refers to whether the analysis of user behavior data has reached the corresponding active index determination.
  • the determination of commentators to be recruited helps to ensure the enthusiasm of the target commentators determined based on the commentators to be recruited to participate in the comment activity.
  • S404 Invoke the voice outbound platform to perform voice interviews for commentators to be recruited, and obtain voice return visit results corresponding to commentators to be recruited.
  • the voice outbound platform is a platform used to implement voice outbound calls for voice interviews.
  • the mobile phone number of the existing user can be obtained, and a one-to-one telephone interview can be conducted with the reviewer user to be recruited by calling the voice outbound platform.
  • the voice return visit result is the result of determining whether a commentator to be recruited meets the commentator recruitment conditions and whether there is a comment willingness to participate in the comment based on the voice return visit.
  • the voice return visit results include the conditions that meet the reviewer recruitment conditions and have the willingness to comment, meet the reviewer recruitment conditions but have no comment willingness, do not meet the reviewer scanning conditions but have the comment willingness, do not meet the reviewer recruitment conditions and have no comment willingness, etc.
  • the server determines that the commentator to be recruited meets the reviewer recruitment conditions and has the willingness to comment based on the voice return visit result of each commentator to be recruited, it will be determined as the target commentator and invited to join the target comment group.
  • Make each target reviewer in the target review group a reviewer who meets the reviewer recruitment conditions and has a willingness to comment, so that the target reviewer is determined through a specific screening, which helps to ensure that the target reviewer Commentators are fixed and their comments are targeted, and the comments made by target commentators on the target comment group are also instantaneous, which helps to ensure the efficiency of software product optimization and shorten the optimization cycle.
  • the target comment group corresponding to the product identifier is created based on the instant messaging tool, so that the creation of the target comment group is targeted, and subsequent targets obtained through the target comment group are guaranteed Comment data is instantaneous. Determining active users who match the product attribute type within a preset time period as commenters to be recruited helps to ensure the enthusiasm of the target commentators determined based on the commentators to be recruited to participate in comment activities.
  • step S403 which is to query the user behavior database based on the product attribute type, and determine the active users matching the product attribute type within a preset time period as commenters to be recruited, including:
  • S501 Query the user behavior database based on the product attribute type, obtain user behavior data within a preset time period before the current time of the system and determine it as user data to be analyzed.
  • Each user data to be analyzed includes a user identification and login duration.
  • the user data to be analyzed is data that meets specific conditions and can be used to analyze and determine the reviewers to be recruited.
  • the server queries the corresponding user behavior database based on the product attribute type, and obtains the user behavior data of all existing users matching the product attribute type; and then filters out that the login time is within a preset time period corresponding to the current time of the system (such as The user behavior data within the past three months) is determined to be user data to be analyzed.
  • the user ID is an ID used to uniquely identify an existing user in a certain software product, such as a mobile phone number.
  • Login duration is the time interval used to log in to a certain software product, which can be understood as the time interval between login time and logout time.
  • S502 Determine the user data to be analyzed whose login duration is greater than the preset duration as valid behavior data.
  • the effective behavior data refers to the behavior data that can be used as an effective analysis to determine the target commentator's corresponding behavior.
  • the user data to be analyzed whose login duration is greater than the preset duration is determined as valid behavior data, which can effectively prevent the user from accidentally clicking the login triggered by the user to interfere with the analysis of the user behavior data, thereby ensuring the accuracy of subsequent data analysis Sex and efficiency.
  • S503 Based on all valid behavior data corresponding to the same user identification, determine the target evaluation index corresponding to the user identification.
  • the target evaluation index is finally determined to be an index that can be used to evaluate whether an existing user is an active user.
  • the target evaluation indicators include, but are not limited to, the number of logins, the number of transactions, and commonly used function information. Specifically, based on all valid behavior data corresponding to the same user identifier, the server statistically analyzes target evaluation indicators such as login times, transaction times, and common function information corresponding to all valid behavior data, so as to determine whether an existing user is an active user.
  • the index evaluation threshold is a preset threshold used to evaluate whether it reaches the active user standard.
  • the evaluation threshold of this indicator includes, but is not limited to, the first threshold used to evaluate whether the number of logins has reached the standard of active users, the second threshold used to evaluate whether the number of transactions has reached the standard of active users, and the second threshold used to evaluate whether commonly used function information is active.
  • User standard breakdown function information In this embodiment, users who have reached at least one of the aforementioned index evaluation thresholds may be determined as commenters to be recruited according to actual needs, so as to ensure the active degree of commentators to be recruited.
  • the user data to be analyzed is user behavior data that matches the product attribute type and the login time is within a preset time period before the current time of the system, which can ensure the finalized comments to be recruited
  • the member is a recent active user
  • the effective behavior data is determined according to the login time, which helps to ensure the accuracy and efficiency of the determined commentators to be recruited
  • the target evaluation index and the index evaluation threshold determined by the effective behavior data are used to determine the commentators to be recruited. To ensure the active level of commentators to be recruited, it can contribute to the enthusiasm of subsequent comments.
  • step S404 that is, calling the voice outbound platform to perform voice interviews with commentators to be recruited, and obtaining the voice return visit results corresponding to commentators to be recruited, including:
  • S601 Obtain the standard interview skills corresponding to the product attribute type, send the standard interview skills and the user ID corresponding to the reviewers to be recruited to the voice outbound platform, and obtain the return visit recording data returned by the voice outbound platform.
  • the standard interview skills are pre-set to guide the commentators to be recruited to reply whether they meet the commentator recruitment conditions and whether they have the willingness to participate in the comments.
  • the server obtains the standard interview skills corresponding to the product attribute type, and sends the standard interview skills and the user identification corresponding to the reviewer to be recruited (such as a mobile phone number or a mobile phone number that can be uniquely determined) to the voice
  • the outbound call platform can perform telephone communication based on the voice outbound platform and the mobile terminal corresponding to the reviewer to be recruited corresponding to the user ID, and obtain the return visit recording data corresponding to the standard interview speech returned by the voice outbound platform.
  • the recording data of the return visit is the voice response of the commentator to be recruited to the standard interview.
  • S602 Use a voice recognition model to recognize the return visit recording data, and obtain return visit text data.
  • the speech recognition model is a pre-trained model for recognizing text content in speech data.
  • the speech recognition model in this embodiment may use a speech static decoding network. Since the static decoding network has fully expanded the search space, the decoding speed is very fast when performing text translation, so that the return visit text data can be quickly obtained.
  • the static speech decoding network is a static decoding network obtained by training with training speech data in a specific field. The training speech data in the specific field can be understood as pre-stored speech data that responds to standard interviews. Since the static speech decoding network is a static decoding network obtained by training based on training speech data in a specific field, it is highly targeted when recognizing the return visit recording data in a specific field, and the decoding accuracy is high.
  • S603 Perform keyword extraction on the return visit text data to obtain return visit keywords.
  • the server uses a keyword extraction algorithm to extract keywords from the return visit text data to obtain return visit keywords.
  • Keyword extraction algorithm refers to an algorithm for extracting keywords from text data.
  • the server first uses the word segmentation tool to segment the return visit text data, and then uses the stop word removal algorithm to perform the stop word removal processing on the word segmentation result to obtain the return visit keywords.
  • S604 Obtain a voice return visit result corresponding to the reviewer to be recruited based on the return visit keyword query evaluation result comparison table.
  • the evaluation result comparison table is a comparison table preset to evaluate whether the reviewer recruitment conditions are met and whether there is a willingness to comment. Understandably, the evaluation result comparison table matches the standard interview technique, that is, the evaluation result comparison table can determine the preset keywords corresponding to different results according to the guiding questions in the standard interview technique, so that the return interview keywords can be compared with Preset keyword matching results and determine the corresponding voice return visit result.
  • the standard interview technique is to set the guiding question of "Are you willing to participate in the project optimization project of software product A”; set the preset keywords corresponding to this guiding question in the evaluation result comparison table as "Yes” and “ “Willing” and “thinking”, etc.; then the return visit keyword extracted from the return visit text data corresponding to this guiding question is "willing", it can be determined that it has a willingness to comment.
  • the reviewer to be recruited is given a voice return visit to obtain the return visit recording data, and the return visit recording data is voice recognized and keyword extraction is performed to determine the return visit keywords.
  • Evaluation result comparison table so as to quickly obtain the voice return interview results corresponding to the reviewers to be recruited, so as to realize the further screening of the reviewers to be recruited, so that the final selected target reviewers are more in line with the review requirements of the software product.
  • a software product evaluation processing device corresponds to the software product evaluation processing method in the above-mentioned embodiment one-to-one.
  • the software product evaluation processing device includes a target comment data acquisition module 701, a target keyword acquisition module 702, a question keyword acquisition module 703, an occurrence frequency acquisition module 704, a high frequency problem determination module 705, and target priority
  • the target comment data acquisition module 701 is used to create a target comment group corresponding to any specific software product installed on the product evaluation platform based on the instant messaging tool, and obtain any target comment in the target comment group
  • the target comment data published by the operator and the corresponding comment time determine whether the target comment data carries a task identifier.
  • the target keyword acquisition module 702 is configured to, if the target comment data does not carry the task identifier, use a keyword extraction algorithm to perform keyword extraction on the target comment data to obtain the target keyword corresponding to the target comment data .
  • the question keyword acquisition module 703 is configured to determine the question keyword corresponding to the target comment data according to the target keyword corresponding to the target comment data and the comment time.
  • the appearance frequency obtaining module 704 is configured to perform frequency statistics on the question keywords in the comment statistical period corresponding to the current time of the system, and obtain the appearance frequency corresponding to the question keywords.
  • the high-frequency question determination module 705 is configured to determine the question keyword as a high-frequency keyword if the occurrence frequency is greater than a preset frequency threshold, and determine the corresponding high-frequency keyword based on the target comment data containing the high-frequency keyword Frequency problem.
  • the target priority obtaining module 706 is configured to obtain the frequency of occurrence of the problem and the degree of impact of the problem corresponding to the high-frequency problem, perform priority analysis based on the frequency of occurrence of the problem and the degree of impact of the problem, and obtain the corresponding Target priority.
  • the response processing module 707 is configured to query the comment response mechanism information table based on the target priority, obtain the corresponding comment response mechanism, and respond to the high-frequency question based on the comment response mechanism.
  • the software product evaluation processing device further includes: a valid comment data determining module 708, configured to obtain the task corresponding to the task identifier if the target comment data carries the task identifier The comment period, when the comment time is within the task comment period, determine the target comment data as valid comment data.
  • the comment tendency result obtaining module 709 is configured to analyze the valid comment data, and obtain the comment question corresponding to the valid comment data and the comment tendency result corresponding to the comment question.
  • the target comment result obtaining module 710 is configured to count the number of comments corresponding to each comment tendency result corresponding to the same comment question associated with the task identifier, and determine the comment tendency result with the largest number of comments as the State the target comment result corresponding to the comment question.
  • the software product evaluation processing device further includes: a historical comment data obtaining unit, a comment participation ratio obtaining unit, a comment available ratio obtaining unit, and a comment reward information obtaining unit.
  • the historical comment data obtaining unit is used to query the comment database based on the user account of the target commentator when the current time of the system is the timed analysis time, and obtain the history of the comment time corresponding to the user account in the data statistical period corresponding to the timed analysis time Comment on the data.
  • the comment participation ratio obtaining unit is configured to query the comment item list corresponding to the data statistical period based on the historical comment data, and obtain the comment participation ratio corresponding to the historical comment data.
  • the comment available ratio obtaining unit is used to obtain the historical tendency result and the target comment result corresponding to the historical comment data, and obtain the comment available ratio corresponding to the historical comment data.
  • the comment reward information obtaining unit is configured to obtain comment reward information corresponding to the user account based on the comment participation ratio and the comment available ratio corresponding to the historical comment data.
  • the question keyword acquisition module 703 includes: a part of speech tagging unit, a keyword determination unit to be analyzed, a first question keyword determination unit, and a second question keyword determination unit.
  • the part-of-speech tagging unit is configured to use a part-of-speech tagging tool to tag the target keywords corresponding to the target comment data, and obtain the part-of-speech corresponding to each target keyword.
  • the keyword determination unit to be analyzed is used to determine the target keyword whose part of speech is the preset part of speech as the keyword to be analyzed.
  • the matching result obtaining unit is configured to use a matching algorithm to perform one-by-one matching processing on the keywords to be analyzed and each preset keyword in the question keyword library to obtain matching results.
  • the first question keyword determining unit is configured to determine the corresponding keyword to be analyzed as the question keyword corresponding to the target comment data if there is a preset keyword whose matching result is a successful match.
  • the second question keyword determining unit is configured to, if there is no preset keyword whose matching result is a successful match, query the database according to the comment time of the target comment data to obtain the comment time before the target comment data.
  • the last piece of historical comment data containing the question keyword, and the question keyword corresponding to the target comment data is determined according to the question keyword of the last piece of historical comment data.
  • the target comment data acquisition module 701 includes: a comment group creation request acquisition unit, a target comment group creation unit, a reviewer to be recruited determination unit, a voice return visit result acquisition unit, and a target reviewer determination unit.
  • the comment group creation request obtaining unit is used to obtain the comment group creation request, and the comment group creation request includes the product identifier and the product attribute type.
  • the target comment group creation unit is used to create a target comment group corresponding to the product identifier based on the instant messaging tool.
  • the reviewer to be recruited determining unit is used to query the user behavior database based on the product attribute type, and determine the active users who match the product attribute type within a preset time period as the reviewer to be recruited.
  • the voice return visit result acquisition unit is used to call the voice outbound platform to perform voice visits to commentators to be recruited, and obtain the voice return visit results corresponding to the commentators to be recruited.
  • the target commentator determination unit is used to determine the commentator to be recruited as the target commentator and invite the target commentator to join the target comment group if the result of the voice return interview meets the commentator recruitment conditions and is willing to comment.
  • the reviewer to be recruited determination unit includes: a user data acquisition subunit to be analyzed, a valid behavior data determination subunit, a target comment indicator determination subunit, and a reviewer to be recruited determination subunit.
  • the user data acquisition sub-unit to be analyzed is used to query the user behavior database based on the product attribute type, and obtain user behavior data within a preset time period before the current time of the system, and determine it as user data to be analyzed.
  • Each user data to be analyzed includes a user identification And login duration.
  • the valid behavior data determining subunit is used to determine the user data to be analyzed whose login duration is greater than the preset duration as valid behavior data.
  • the target comment index determination subunit is used to determine the target evaluation index corresponding to the user identifier based on all effective behavior data corresponding to the same user identifier.
  • the reviewer-to-be-recruited determining subunit is used to determine the user corresponding to the user identifier as the reviewer-to-be-recruited if the target evaluation index meets the corresponding index evaluation threshold.
  • the voice return visit result acquisition unit includes: a return visit recording data acquisition subunit, a return visit text data acquisition subunit, a return visit keyword acquisition subunit, and a return visit result acquisition subunit.
  • the return visit recording data acquisition sub-unit is used to obtain the standard interview skills corresponding to the product attribute type, send the standard interview skills and the user ID corresponding to the reviewers to be recruited to the voice outbound platform, and obtain the return from the voice outbound platform Revisit the recording data.
  • the return visit text data acquisition subunit is used to recognize the return visit recording data using a voice recognition model and obtain the return visit text data.
  • the return visit keyword acquisition subunit is used to extract keywords from the return visit text data to obtain return visit keywords.
  • the return visit result acquisition sub-unit is used to obtain the voice return visit result corresponding to the reviewer to be recruited based on the search evaluation result comparison table of the return visit keyword.
  • each module in the above-mentioned software product evaluation processing device can be implemented in whole or in part by software, hardware, and combinations thereof.
  • the foregoing modules may be embedded in the form of hardware or independent of the processor in the computer device, or may be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the foregoing modules.
  • a computer device is provided.
  • the computer device may be a server, and its internal structure diagram may be as shown in FIG. 8.
  • the computer equipment includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide calculation and control capabilities.
  • the memory of the computer device includes a non-volatile storage medium and an internal memory.
  • the non-volatile storage medium stores an operating system, a computer program, and a database.
  • the internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium.
  • the database of the computer equipment is used to store the data collected and generated during the execution of the software product evaluation processing method, such as the target comment result.
  • the network interface of the computer device is used to communicate with an external terminal through a network connection.
  • the computer program is executed by the processor to realize a software product evaluation processing method.
  • a computer device including a memory, a processor, and a computer program stored in the memory and capable of running on the processor.
  • the processor executes the computer program to implement the software product evaluation processing method in the foregoing embodiment , Such as S201-S210 shown in Figure 2, or shown in Figures 3 to 6.
  • the function of each module/unit in the embodiment of the software product evaluation processing device is realized, for example, the function of each module shown in FIG. 7.
  • a computer-readable storage medium is provided, and a computer program is stored on the computer-readable storage medium.
  • the computer program is executed by a processor, the software product evaluation processing method in the above-mentioned embodiment is implemented, as shown in FIG. 2 S201-S210, or as shown in Figures 3 to 6.
  • Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
  • Volatile memory may include random access memory (RAM) or external cache memory.
  • RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
  • SRAM static RAM
  • DRAM dynamic RAM
  • SDRAM synchronous DRAM
  • DDRSDRAM double data rate SDRAM
  • ESDRAM enhanced SDRAM
  • SLDRAM synchronous chain Channel
  • memory bus Radbus direct RAM
  • RDRAM direct memory bus dynamic RAM
  • RDRAM memory bus dynamic RAM

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Accounting & Taxation (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Finance (AREA)
  • Computational Linguistics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

涉及金融工具技术领域，一种软件产品测评处理方法、装置、计算机设备及存储介质。该方法包括：基于即时通讯工具创建目标评论群，获取目标评论数据和对应的评论时间(S201)；若目标评论数据未携带任务标识，则采用关键词提取算法进行关键词提取，获取目标关键词(S202)；根据目标关键词和评论时间，确定问题关键词(S203)；对系统当前时间对应的评论统计周期内的问题关键词进行频次统计，获取出现频次(S204)；若出现频次大于预设频次阈值，则确定高频关键词和对应的高频问题(S205)；对高频问题进行优先级分析，获取目标优先级(S206)；基于目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于评论响应机制对高频问题进行响应处理(S207)。该方法可提高软件产品测评效率。

Description

软件产品测评处理方法、装置、计算机设备及存储介质
本申请要求于2019年5月22日提交中国专利局、申请号为201910430101.5，发明名称为“软件产品测评处理方法、装置、计算机设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及金融工具技术领域，尤其涉及一种软件产品测评处理方法、装置、计算机设备及存储介质。
背景技术
当前银行、证券和保险等金融机构或者其他产品开发机构开发不同的软件产品(如银行机构开发的理财产品APP)，以便基于该软件产品进行业务推广。在软件产品上线之后，需进行测评，以便基于测评结果进一步优化软件产品，即需要采集用户的评论数据，对用户的评论数据进行分析，以便根据分析结果对软件产品进行优化，从而吸引更多用户使用软件产品。当前软件产品测评时，采用用户访谈或者调查问卷等传统调研方式，需要产品开发机构邀请相关评论员在测评周期内对自主设置的测评问题进行评论，获取评论数据并进行分析，得到分析结果。发明人意识到，这种软件产品测评过程存在如下几点不足：其一是，成本较高、效率较低，需要由产品开发机构配备相应的人员进行测评问题设计和后续测评分析，其人力成本和时间成本较高，且效率较低；其二是，实时性不强，一般只采集评论员在测评周期内的评论数据，无法实时采集评论员在使用软件产品过程的想法或问题，进而进行产品优化；其三是，分析结果具有局限性，由于评论数据一般是针对自主设置的测评问题进行评论，使其分析结果围绕测评问题，无法及时反映评论员使用软件产品过程中的想法。
发明内容
本申请实施例提供一种软件产品测评处理方法、装置、计算机设备及存储介质，以解决当前软件产品测评过程中存在的成本较高、效率较低、实时性不强和分析结果具有局限性的问题。
一种软件产品测评处理方法，包括：
基于即时通讯工具创建安装产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识；
若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词；
根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词；
对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词对应的出现频次；
若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题；
获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级；
基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
一种软件产品测评处理装置，包括：
目标评论数据获取模块，用于基于即时通讯工具创建安装在产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识；
目标关键词获取模块，用于若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词；
问题关键词获取模块，用于根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词；
出现频次获取模块，用于对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词对应的出现频次；
高频问题确定模块，用于若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题；
目标优先级获取模块，用于获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级；
响应处理模块，用于基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现上述软件产品测评处理方法。
一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现上述软件产品测评处理方法。
上述软件产品测评处理方法、装置、计算机设备及存储介质，通过即时通讯工具创建与特定软件产品相对应的目标评论群，并获取目标评论群中任一目标评论员发表的目标评论数据，从而保证目标评论数据处理的实时性。
附图说明
图1是本申请一实施例中软件产品测评处理方法的一应用环境示意图；
图2是本申请一实施例中软件产品测评处理方法的一流程图；
图3是本申请一实施例中软件产品测评处理方法的另一流程图；
图4是本申请一实施例中软件产品测评处理方法的另一流程图；
图5是本申请一实施例中软件产品测评处理方法的另一流程图；
图6是本申请一实施例中软件产品测评处理方法的另一流程图；
图7是本申请一实施例中软件产品测评处理装置的一示意图；
图8是本申请一实施例中计算机设备的一示意图。
具体实施方式
本申请实施例提供的软件产品测评处理方法，该软件产品测评处理方法可应用如图1所示的应用环境中。具体地，该软件产品测评处理方法应用在产品测评平台中，该产品测评平台是用于实现对软件产品进行测评的综合平台。该产品测评平台包括如图1所示的客户端和服务器，客户端与服务器通过网络进行通信，用于实现对软件产进行调研分析，既保证评论员评论的针对性和评论时间的灵活性，以便根据最终获取的目标评论结果进行软件产品优化，有效降低软件产品优化周期。其中，客户端又称为用户端，是指与服务器相对应,为客户提供本地服务的程序。客户端可安装在但不限于各种个人计算机、笔记本电脑、智能手机、平板电脑和便携式可穿戴设备上。服务器可以用独立的服务器或者是多个 服务器组成的服务器集群来实现。
如图2所示，提供一种软件产品测评处理方法，应用在产品测评平台上，该方法应用在图1所示的服务器上，具体包括如下步骤：
S201：基于即时通讯工具创建安装在产品测评平台上的任一特定软件产品对应的目标评论群，获取目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断目标评论数据是否携带任务标识。
其中，目标评论群是基于即时通讯工具创建与用于对特定软件产品进行评论的即时通信群，该目标评论群是由系统管理员和目标评论员组成的用于对特定软件产品进行评论的即时通信群。本实施例中，特定软件产品安装在产品测评平台上，以使该产品测评平台可以对该特定软件产品进行测评，保障软件产品综合测评的可执行性，无需产品开发机构配置相应的人员进行测评处理，节省人力成本。该系统管理员是用于创建并管理目标评论群的群成员，主要用于发表评论任务或者收集评论数据，一般为产品开发机构内部的工作人员。目标评论员是目标评论群中的群成员，更具体地，该目标评论员具体为满足预先设置的评论员条件的群成员，主要用于发表评论数据，使得发表目标评论数据的目标评论员具有固定性，节省目标评论员的确定时间，有利于保障目标评论数据的针对性。该预先设置的评论员条件具体可以包括参与活动的活跃程度和是否有意向等条件。每一目标评论数据对应一评论时间，该评论时间是目标评论员在目标评论群上发表目标评论数据的时间。本实施例中，目标评论数据是目标评论员在目标评论群上发表的数据，有助于保障评论数据分析的时效性。
具体地，服务器获取目标评论群中任一目标评论员发表的目标评论数据，具体包括如下两种情况：
第一种，具体包括如下步骤：1-1)服务器获取系统管理员通过评论终端发表的待评论任务，并在客户端的目标评论群上显示待评论任务，待评论任务包括任务标识、任务内容和任务评论期限。1-2)服务器获取目标评论群中任一目标评论员通过评论终端发表的采用特定评论数据格式的与待评论任务相对应的目标评论数据，目标评论数据对应一评论时间。这种情况下，采用特定评论数据格式的目标评论数据中携带任务标识。
其中，评论终端是用于供系统管理员或者目标评论员使用的用于在目标评论群上发表信息的终端，可以为安装有即时通信客户端的终端。该待评论任务是基于产品开发机构主导的专题问题所创建的需要评论的任务，即该待评论任务是针对产品开发机构预先设计的调研问题发布的任务。任务标识是用于区别于某一评论任务区别于其他评论任务的标识，可以按照预设的任务编号规则进行设置，该任务编号规则可以采用任务标签和顺序标签来确定，例如RW001和RW002中，RW为任务标签，001和002为顺序标签。一般来说，每一任务标识对应一待评论任务。任务内容是用于说明本次待评论任务所要针对的评论主题，可以为如某一功能的使用感受、痛点或者对新功能的想法等。任务评论期限是用于限定本次待评论任务的评论期限，以促使目标评论员及时评论，以保证软件产品优化的效率。本实施例中，可由系统管理员将该待评论任务发表在目标评论群中，以使目标评论群中所有目标评论员均可对该待评论任务进行评论，以获取特定评论数据格式的目标评论数据。
本实施例中，为了保证目标评论员所发表的目标评论数据与该待评论任务相关联，可预先配置特定评论数据格式，以使目标评论员采用特定评论数据格式发表相应的目标评论数据。在一实施例中，该特定评论数据格式可以为：评论标识(如#)-任务标识(如RW001或RW002)-评论内容数据。可以理解地，目标评论员采用这种特定评论数据格式发表目标评论数据，可有助于对同一评论标识相对应的目标评论数据进行统一分析处理，提高软件产品的调研效率，进而缩短软件产品优化周期。
进一步地，为了保证评论终端所发表的目标评论数据均在任务评论期限内，并方便采用特定评论数据格式上传目标评论数据，可以在目标评论员通过评论终端输入相应的评论 标识(如#)之后，在该评论终端上以列表形式显示当前处于任务评论期限内的所有待评论任务的待评论任务列表，以便目标评论员基于待评论任务列表选择需要进行评论的待评论任务，该待评论任务列表上显示每一任务标识和对应的任务内容。目标评论员选择待评论任务列表中的任一待评论任务之后，再通过评论终端输入评论内容数据，即可获取“评论标识(如#)-任务标识(如RW001或RW002)-评论内容数据”这种特定评论数据格式的目标评论数据，以保证特定评论数据格式的目标评论数据的发表效率。
第二种，具体包括如下步骤：2-1)服务器获取目标评论群中任一目标评论员基于评论终端随机发表的未采用特定评论数据格式的目标评论数据。即这种情况下，未采用特定评论数据格式的目标评论数据中未携带任务标识。
本实施例中，目标评论员在使用该特定软件产品过程中产生的新想法或者认为存在某种问题或痛点，对软件产品功能改进可能有用时，可以通过评论终端发表未采用特定评论数据格式的目标评论数据。该目标评论数据可以为对该特定软件产品已有功能的问题类反馈，也可以是自主提出的新的需求类反馈，也即该目标评论数据不是针对系统管理员已经发表的待评论任务进行评论的数据。可以理解地，可通过分析目标评论群中任一目标评论员随机发表的目标评论数据，使其评论时间具有灵活性，有助于归纳分析出可进行软件产品优化改进的新的功能规划或新想法，并对新的功能规划和新想法进行优化完善，以便进行软件产品优化，避免直接优化的软件产品存在问题而需要反复修改，从而缩短软件产品的优化周期。
具体地，服务器获取目标评论群中任一目标评论员发表的目标评论数据之后，可采用字符串匹配算法或者正则表达式匹配算法对目标评论数据进行处理，以判断目标评论数据中是否携带任务标识。如上述实施例中，若在目标评论数据中匹配到RW这一任务标签和顺序标签时，即可认定该目标评论数据携带任务标识。
S202：若目标评论数据未携带任务标识，则采用关键词提取算法对目标评论数据进行关键词提取，获取目标评论数据对应的目标关键词。
其中，目标关键词是对目标评论数据进行关键词提取之后，所确定的可以用于反映目标评论数据的关键词。关键词提取算法是用于提取文本中关键词的算法，可以采用但不限于TextRank算法，或者可以通过对目标评论数据进行分词和去停用词等处理，以确定其对应的目标关键词。具体地，若目标评论数据未携带任务标识，则说明该目标评论数据不是基于系统管理员预先发布的待评论任务所发表的评论，而是目标评论员在日常使用软件产品过程中产生的新想法、痛点或者问题等内容反馈的评论。进一步地，若目标评论数据未携带任务标识时，其所发表的目标评论数据可以为问题类反馈也可以为需求类反馈。其中，问题类反馈是指对软件产品的现有功能、产品和操作流程过程中出现的问题的反馈。需求类反馈是指目标评论员期望该软件产品需增加的新功能的反馈。
本实施例中，若目标评论数据未携带任务标识，则服务器采用关键词提取算法对目标评论员发表的目标评论数据进行关键词提取，以提取出可反映目标评论数据中关键内容的目标关键词，以便后续进行数据的统计分析处理，提高处理效率。
S203：根据目标评论数据对应的目标关键词和评论时间，确定目标评论数据对应的问题关键词。
由于目标评论数据未携带任务标识时，说明目标评论员所发表的这一目标评论数据不是针对系统管理员预先发布的待评论任务的评论，此时目标评论数据所针对的评论问题无法直接确定，需分析目标评论数据的具体内容，以确定这一条目标评论数据所针对的评论问题。目标评论数据对应的问题关键词具体是指该目标评论数据所针对的评论问题对应的关键词。例如，“这个APP的登录界面容易卡顿，体验不好”这一个目标评论数据中，问题关键词可以为“登录界面”，即该问题关键词主要是针对特定软件产品中某一功能界面，以确定评论主题是针对这一功能界面的问题。
具体地，步骤S203具体包括如下步骤：
S2031：采用词性标注工具对所述目标评论数据对应的目标关键词进行词性标注，获取每一所述目标关键词对应的词性。其中，词性标注工具是用于对任一目标关键词的词性进行标注的工具，包括但不限于PosTagger工具。
S2032：将词性为预设词性的目标关键词确定为待分析关键词。其中，预设词性是系统预先设置的需要进行分析的词性，如名词、动词和形容词。
S2033：采用匹配算法对所述待分析关键词和问题关键词库中的每一预设关键词进行逐一匹配处理，获取匹配结果。其中，问题关键词库为预先设置的用于存储预设关键词的数据库。该预设关键词为预先设置的与评论问题相关的关键词。匹配算法为预先设置的用于匹配两者是否相同或者相似的算法，包括但不限于正则表达式匹配算法。例如，若采用正则表达式匹配算法匹配到与待分析关键词相同的预设关键词，则获取的匹配结果为匹配成功，反之，获取的匹配结果为匹配失败。
S2034：若存在所述匹配结果为匹配成功的预设关键词，则将对应的待分析关键词确定为所述目标评论数据对应的问题关键词。
例如，“这个APP的登录界面容易卡顿，体验不好”这一目标评论数据中，匹配出该目标评论数据所提取的目标关键词包含问题关键词“登录界面”和“卡顿”等。
S2035：若不存在所述匹配结果为匹配成功的预设关键词，则根据所述目标评论数据的评论时间查询数据库，获取所述目标评论数据的评论时间之前的包含问题关键词的最近一条历史评论数据，根据最近一条所述历史评论数据的问题关键词确定所述目标评论数据对应的问题关键词。
例如，“我也碰到这个问题”或者“同意上述观点”等没有明确指示其对应的评论问题的目标评论数据中，所识别的待分析关键词中不存在匹配结果为匹配成功的预设关键词，此时，可根据目标评论数据的评论时间查询数据库，获取该目标评论数据的评论时间之前的包含问题关键词的最近一条历史评论数据，如“这个APP的登录界面容易卡顿，体验不好”这一历史评论数据，则直接将这一历史评论数据对应的问题关键词确定为该目标评论数据的问题关键词。
可以理解地，这种在目标评论数据中明确描述其评论问题时，可直接确定问题关键词；在目标评论数据中没有明确描述其评论问题时，可将目标评论数据的评论时间之前最近一个历史评论数据的问题关键词确定为该目标评论数据的问题关键词，利用目标评论数据的评论的连续性，有助于统计问题发生频率，保障后续分析确定高频问题的客观性。
S204：对评论时间在系统当前时间对应的评论统计周期内的问题关键词进行频次统计，获取问题关键词对应的出现频次。
其中，系统当前时间是指服务器的系统时间。评论统计周期是预先设置的用于对目标评论数据对应的问题关键词进行统计分析的周期，该评论统计周期可以为一周或者其他时间。系统当前时间对应的评论统计周期是指从系统当前时间之前与评论统计周期相对应的时间区间，即以系统当前时间为统计的截止时间，以保证问题关键词的出现频次统计的实时性。例如，若系统当前时间为2月10日，而评论统计周期为3天，则系统当前时间对应的评论统计周期为2月8日、2月9日和2月10日。
由于每一目标评论数据对应一评论时间，而从目标评论数据中提取出的问题关键词也对应该评论时间。本实施例中，对评论时间在系统当前时间对应的评论统计周期内的问题关键词进行频次统计，获取问题关键词对应的出现频次，具体包括：服务器先获取评论时间在系统当前时间对应的评论统计周期内的所有目标评论数据，对该目标评论数据包含该问题关键词的次数进行统计，以获取该问题关键词对应的出现频次。
S205：若出现频次大于预设频次阈值，则将问题关键词确定为高频关键词，基于包含高频关键词的目标评论数据确定对应的高频问题。
其中，预设频次阈值是预先设置的用于评估是否达到认定为高频率出现的次数的阈值。具体地，服务器将问题关键词对应的出现频次与预先设置的预设频次阈值进行比较，若该问题关键词对应的出现频次大于预设频次阈值，则认定该问题关键词为高频关键词，将包含该高频关键词对应的目标评论数据中提取出现的评论问题确定为高频问题。本实施例中，通过比较问题关键词在系统当前时间对应的评论统计周期内的出现频次与预设频次阈值的大小，确定该问题关键词是否为高频关键词，基于包含该高频关键词的目标评论数据，确定其对应的高频问题，以保证高频问题确定的时效性(既需在系统当前时间对应的评论统计周期内的数据)，有助于提高软件产品的优化效率。
本实施例中，基于包含高频关键词的目标评论数据确定对应的高频问题具体包括：若包含高频关键词的目标评论数据只有一条，则直接根据所述目标评论数据中包含的
S206：统计高频问题对应的问题发生频率和问题影响程度，依据问题发生频率和问题影响程度进行优先级分析，获取高频问题对应的目标优先级。
其中，问题发生频率是依据该高频问题在目标评论群中的响应情况确定的出现频率的分析结果。具体地，服务器在认定某一评论问题为高频问题后，在目标评论群上发布相应的信息响应任务，获取目标评论群中所有目标评论员对该信息响应任务的问题响应信息，基于该问题响应信息统计问题发生概率，并将该问题发生概率与预设概率阈值进行比较，若问题发生概率大于预设概率阈值，则获取问题发生概率较大的问题发生概率结果；若问题发生概率不大于预设概率阈值，则获取问题发生概率较小的问题发生概率结果。其中，问题发生频率较小出现在较少目标评论员遇到这一高频问题的情况下，而问题发生频率较大出现在较多目标评论员遇到这一高频问题的情况下。信息响应任务是系统管理员发布的用于采集目标评论员是否遇到某一高频问题的任务。问题响应信息是目标评论中对信息响应任务进行响应的信息。预设概率阈值是预先设置的与概率相关的阈值。
例如，在1000个目标评论员的目标评论群中，目标评论员A提出一目标评论数据并认定其为高频问题，系统管理员基于该高频问题发表信息响应任务时，若获取的问题响应信息为有100个目标评论员响应自己也曾遇到过相应的问题，基于该问题响应信息统计的问题发生概率为10％，将该问题发生概率与预设概率阈值(如5％)进行比较，即可确定其对应的问题发生频率为问题发生概率较大。
其中，问题影响程度是用于反映高频问题对用户体验影响的程度指标。该问题影响程度可根据问题的严重性等级确定。本实施例中，问题的严重性等级采用Wilson(1990)的五级标准如下：5级-无关紧要的错误；4级-问题虽小但却让用户焦躁；3级-中等程度，耗费时间但不会丢失数据；2级-导致数据丢失的严重问题；1级-灾难性错误，导致数据的丢失或者软硬件的损坏。或者，问题的严重性等级可采用另一五级标准如下：0-不认为是一个可用性问题；1-这仅仅是一个装饰门面的可用性问题：并不需要特别的处理，除非这个项目有额外的时间；2-次要的(Minor)可用性问题：解决这个问题的优先级较低；3-主要的(Major)可用性问题：解决这个问题是很重要的，优先级很高；4-可用性灾难(Catastrophe)：解决这个问题是非常必要而且紧急的(Imperative)，必须在软件产品发布上线之前解决。本实施例中，问题影响程度与问题的严重性等级相匹配，如包括五级的问题影响程度。或者，问题影响程度可根据问题的严重性等级计算确定，如将五级的严重性等级划分成二级的问题影响程度，如Wilson(1990)的五级标准中，第5级和第4级标准的问题确定为问题影响程度较小的问题影响程度，第1、2和3级标准的问题确定为问题影响程度较大的问题影响程度。
本实施例中，对高频问题进行优先级分析，获取高频问题对应的目标优先级，具体包括：对高频问题进行优先级分析，获取高频问题对应的问题发生频率和问题影响程度；基于问题发生频率和问题影响程度查询优先级信息对照表，获取高频问题对应的目标优先级。其中，优先级信息对照表是用于反映不同的问题发生频率和问题影响程度组合的优先 级信息的对照表(如下表一所示)。
具体地，对高频问题进行优先级分析，是指基于预先设置的评估指标判断标准，确定高频问题对应的问题发生频率和问题影响程度这两种评估指标进行综合分析，以确定其对应的目标优先级的过程。目标优先级是指基于问题发生频率和问题影响程度进行综合分析后确定的优先级，该目标优先级在一定程度上反馈高频问题的紧急程度。
表一优先级信息对照表
S207：基于目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于评
Figure PCTCN2020085579-appb-000001
论响应机制对高频问题进行响应处理。
其中，评论响应机制信息表是用于反映每一种优先级对应的高频问题的响应机制的信息表。其中，响应机制是对某一高频问题进行响应的响应时间和响应流程进行规定的机制。响应时间是指对高频问题进行响应的时间。响应流程具体可以包括对某一高频问题配置的进行响应的流程和对应的处理人员，以使对高频问题的响应处理更合理和更高效，确保高频问题的处理效率，提高软件产品的优化效率，缩短软件产品的优化周期。一般来说，目标优先级越先，表明该高频问题越紧急，越需要优先处理，则其响应时间越短，其配置的响应流程越便捷，其参与的处理人员越多或者主管人员的岗位越高级等。
本实施例所提供的软件产品测评处理方法中，通过即时通讯工具创建创建安装在产品测评平台上的任一特定软件产品相对应的目标评论群，并获取目标评论群中任一目标评论员发表的目标评论数据，从而保证目标评论数据处理的实时性。在目标评论数据未携带任务标识时，说明该目标评论数据不是针对待评论任务发表的评论，使得目标评论数据的评论内容更具有灵活性，避免局限性。然后，采用关键词提取算法对目标评论数据进行关键词提取，获取目标关键词，以保证后续数据处理的效率。根据目标评论数据对应的目标关键词和评论时间，确定对应的问题关键词，从而确定每一目标评论数据对应的评论问题，保障后续分析确定高频问题的客观性。通过比较问题关键词在系统当前时间对应的评论统计周期内的出现频次与预设频次阈值的大小，确定该问题关键词是否为高频关键词，从而确定高频问题，以保证高频问题确定的时效性，有助于提高软件产品的优化效率。最后，依据问题发生频率和问题影响程度对高频问题进行优先级分析，以确定对应的目标优先级之后，可根据目标优先级确定的评论响应机制对高频问题进行响应处理，以使对高频问题的响应处理更合理和更高效，确保高频问题的处理效率，提高软件产品的优化效率，缩短软件产品的优化周期。
在一实施例中，在步骤S201中的判断所述目标评论数据是否携带任务标识之后，该软件产品测评处理方法还包括：
S208：若目标评论数据携带任务标识，则获取任务标识对应的任务评论期限，在评论时间在任务评论期限内时，将目标评论数据确定为有效评论数据。
其中，若目标评论数据携带任务标识，则说明该目标评论数据是基于系统管理员预先发布的待评论任务所发表的评论，所针对的评论问题与待评论任务中的任务内容相匹配，可以理解为针对任务内容进行的评论。具体地，若目标评论数据携带任务标识，则服务器需获取目标评论数据对应的评论时间，并基于任务标识查询待评论任务列表，获取与该任务标识相对应的任务评论期限，判断该目标评论数据的评论时间是否在该任务评论期限内，以便分析该目标评论数据是否有效，从而保障待评论任务的时效性。其中，待评论任 务列表是用于记录所有待评论任务的任务信息的数据表，具体用于存储每一待评论任务对应的任务标识、任务内容和任务评论期限等任务信息。
具体地，若目标评论数据的评论时间在该任务标识对应的任务评论周期内，则说明目标评论员是在该任务标识对应的待评论任务的任务评论周期内发表的评论，是对待评论任务的及时响应，可以将该目标评论数据作为后续数据分析的依据，即有效评论数据。其中，有效评论数据是指评论时间在对应的任务评论周期内的可作为数据分析依据的评论数据。
S209：对有效评论数据进行分析，获取有效评论数据对应的评论问题和与评论问题相对应的评论倾向结果。
其中，服务器对有效评论数据进行分析，是指用于分析每一有效评论数据所针对的评论问题以及该目标评论员对这一评论问题有什么倾向的分析过程。具体地，服务器可采用自然语言处理工具对有效评论数据进行分析，以获取有效评论数据对应的评论问题和与评论问题相对应的评论倾向结果。
评论问题是指该有效评论数据所针对的问题，该评论问题可以与待评论任务中的任务内容相匹配。例如，若有效评论数据中明确提及所针对的评论问题时，可直接确定有效评论数据对应的评论问题；若有效评论数据中没有明确提及所针对的评论问题(如采用省略问题而直接引用待评论任务的形式发表的有效评论数据)，可依据待评论任务中的任务内容确定对应的评论问题。
评论倾向结果是从目标评论员发表的有效评论数据中提取其对某一评论问题的某一种倾向的结果。例如，若有效评论数据是针对特定软件产品的数据上传功能是否好用进行评论这一评论问题进行评论的数据，则评论倾向结果包括好用和不好用两个不同倾向。若有效评论数据是针对P1、P2和P3这三种版本的数据上传功能，评价哪种更好用的评论问题，则其评论倾向结果包括P1、P2或者P3更好用的评论倾向结果。
S210：统计与任务标识相关联的同一评论问题对应的每一评论倾向结果对应的评论数量，将评论数量最多的评论倾向结果确定为评论问题对应的目标评论结果。
具体地，服务器统计该目标评论群中基于同一任务标识对应的待评论任务对应的有效评论数据中，同一评论问题对应的每一评论倾向结果对应的评论数量，该评论数量可以理解为目标评论群上支持该评论倾向结果的目标评论员的数量。然后，服务器将评论数量最多的评论倾向结果确定为该评论问题对应的目标评论结果。
例如，若有效评论数据是针对P1、P2和P3这三种版本的数据上传功能，评价哪种更好用的评论问题，不同目标评论员有不同的评论倾向结果，服务器需要对这一评论问题的每一种评论倾向结果进行统一汇总，以获取每一种评论倾向结果对应的评论数量，以将评论数量最多的评论倾向结果确定为该评论问题对应的目标评论结果，即采用多数原则确保目标评论结果的客观性。例如，在1000条有效评论数据中，认为P1、P2和P3这三种版本的数据上传功能更好用的评论数量分别为200、300和500，则将P3版本的数据上传功能更好用确定为该评论问题对应的目标评论结果。
本实施例所提供的软件产品测评处理方法中，通过获取目标评论群上任一目标评论员发表的目标评论数据，以使发表目标评论数据的评论员具有固定性，有利于保障目标评论数据的针对性。在目标评论数据携带任务标识时，若目标评论数据对应的评论时间在任务标识对应的任务评论期限内，则将目标评论数据确定为有效评论数据，以保证有效评论数据的评论时效性，有助于提高评论数据处理的时效性，从而缩短软件产品的优化周期。在对有效评论数据进行分析确定评论问题和评论倾向结果之后，统计同一评论问题对应的每一评论倾向结果的评论数量，以将评论数量最多的评论倾向结果确定为该评论问题对应的目标评论结果，以使目标评论结果更具有客观性，有助于保障软件产品优化的质量。
在一实施例中，为了鼓励目标评论群中的目标评论员及时发表更多对软件产品优化具有建设性意义的目标评论数据，可以设置不同的奖励标准对每一目标评论员的评论行为进 行分析处理，并根据定时分析处理的结果提供相应的奖励。可以理解地，服务器可预先设置相应的定时分析任务，用于实现对每一目标评论员的评论行为进行定时分析。该定时分析任务是预先设置的用于在某一定时分析时间执行相应的分析过程的任务。其中，定时分析时间是预先设置的用于进行分析处理的时间。如图3所示，在步骤S210之后，即在将评论数量最多的评论倾向结果确定为评论问题对应的目标评论结果之后，软件产品测评处理方法还包括：
S301：在系统当前时间为定时分析时间时，基于目标评论员的用户帐号查询评论数据库，获取评论时间在定时分析时间对应的数据统计周期内的与用户帐号相对应的历史评论数据。
其中，历史评论数据是指在系统当前时间之前该目标评论员所发表的可用于进行数据统计分析的目标评论数据。目标评论员的用户帐号是指每个目标评论员在该目标评论群中的唯一帐号，可用于区分不同目标评论员发表的目标评论数据。评论数据库是用于存储目标评论库中所有目标评论员发表的目标评论数据的数据库。可以理解地，存储在评论数据库中的每一目标评论数据与一用户帐号关联存储，以便对每一目标评论员的评论行为进行综合分析。
数据统计周期是预先设置的用于对任一目标评论员的目标评论数据进行统计分析的周期。定时分析时间对应的数据统计周期是指在定时分析时间之前与数据统计周期相对应的时间区间。例如，定时分析时间为每周一12：00，而数据统计周期为一周，则定时分析时间对应的数据统计周期是指从定时分析时间之前的一周。
本实施例中，步骤S301具体包括：在系统当前时间为定时分析时间时，服务器执行预先设置的定时分析任务，即先基于每一目标评论员的用户帐号查询评论数据库，获取该用户帐号对应的所有目标评论数据并确定每一目标评论数据对应的评论时间；然后，将评论时间在定时分析时间对应的数据统计周期内的目标评论数据，确定为该用户帐号对应的可用于进行数据统计分析的历史评论数据。
S302：基于历史评论数据查询与数据统计周期相对应的评论项目列表，获取历史评论数据对应的评论参与比例。
其中，与数据统计周期相对应的评论项目列表是用于记录该数据统计周期内所有评论项目的信息表。该评论项目包括但不限于系统管理员自主确定的待评论任务对应的项目，还包括基于高频问题所创建的新增评论任务对应的项目。评论参与比例是用于反馈目标评论员参与评论的参与项目数量占整个评论项目列表中所有项目的项目总数的比值。例如，若在一数据统计周期内的评论项目有20个，即项目总数为20，而一目标评论员参与评论的参与项目数量为16个，则获取的评论参与比例为80％。可以理解地，该评论参与比例越高，说明该目标评论员越积极参与对评论项目的评论，可以客观地反映目标评论员在目标评论群中的活跃程度。
S303：获取历史评论数据对应的历史倾向结果和目标评论结果，获取历史评论数据对应的评论可用比例。
具体地，对于待评论任务对应的项目，每一目标评论数据已经预先进行过分析，可直接将历史评论数据对应的评论倾向结果作为历史倾向结果，相应地，并可直接获取评论问题对的目标评论结果。对于新增评论任务对应的项目，可对历史评论数据进行分析，以确定该历史评论数据对应的评论问题和与评论问题相对应的历史倾向结果，其过程如步骤S204所示，为避免重复，在此不一一赘述；并参考步骤S205确定该评论问题对应的目标评论结果。
其中，评论可用比例是指历史倾向结果与目标评论结果相匹配的历史评论数据的匹配数量与所有历史评论数据对应的参与项目数量的比例。例如，若在一数据统计周期内，一目标评论员对应的所有历史评论数据对应的参与项目数量为16，其中，有12个评论项目 的历史倾向结果与最终确定的目标评论结果相同，即匹配数量为12，则计算出的评论可用比例为75％。可以理解地，该评论可用比例越高，说明目标评论员的评论建议对于软件产品优化的作用越大，即其评论有用性越高。
S304：基于历史评论数据对应的评论参与比例和评论可用比例，获取用户帐号对应的评论奖励信息。
其中，步骤S304具体包括如下步骤：(1)服务器基于历史评论数据对应的评论参与比例和评论可用比例查询预先设置分值对照表，分别确定该评论参与比例对应的参与分值P1和评论可用比例对应的可用分值P2。其中，分值对照表是预先设置的用于规定每一评论参与比例与对应的参与分值的相互关系，和评论可用比例与对应的可用分值的相互关系的对照表。(2)再获取该评论参与比例对应的参与权重W1和评论可用比例对应的可用权重W2，其中，W1+W2＝1。为了更有效地提高目标评论员参与评论的积极性，可设置该参与权重W1大于可用权重W2，如W1＝70％，而W2＝30％。(3)采用加权算法P＝P1*W1+P2*W2计算该目标评论员对应的评论分值P。本实施例中，P1和P2采用同一数值范围的分值，以使两者具有可比值，如均为0-10之间的数值，使得计算出来的评论分值也为0-10之间的数值。(4)基于评论分值P查询评论奖励标准表，获取与该评论分值P相对应的评论奖励信息，确定为该用户帐号对应的评论奖励信息。该评论奖励信息包括但不限于现金红包奖励、会员特权类奖励以及实体礼品奖励，可根据不同评论分值P自主确定。可以理解地，评论分值P越大，其对应的评论奖励信息中奖励越丰盛，越有助于吸引目标评论员积极发表目标评论数据，调动目标评论员的积极性。该评论奖励信息的确定是基于历史评论数据对应的评论参与比例和评论可用比例进行综合分析计算后确定的，使其具有客观性。
本实施例所提供的软件产品测评处理方法中，通过目标评论员的用户帐号查询评论数据，以确定与数据统计周期相对应的历史评论数据，以使基于历史评论数据进行评论行为分析具有周期性，更有利于促使目标评论员积极发表目标评论数据。在对历史评论数据进行统计分析过程中，确定对应的评论参与比例和评论可用比例，以分别从活跃程度和评论有用性这两个维度客观地对目标评论员进行评估。再基于历史评论数据对应的评论参与比例和评论可用比例，获取相应的评论奖励信息，以保证评论奖励信息的客观性，更有助于吸引目标评论员积极地在目标评论群上发表目标评论数据。
如图4所示，在步骤S201中，基于即时通讯工具创建安装所述产品测评平台上的任一特定软件产品对应的目标评论群，具体包括如下步骤：
S401：获取评论群创建请求，评论群创建请求包括产品标识和产品属性类型。
其中，评论群创建请求是用于触发对产品测评平台上的某一特定软件产品创建评论群的请求。产品标识是用于唯一区别不同软件产品的标识。产品属性类型是用于限定本次所要创建的评论群需要评论的软件产品的属性。该产品属性类型包括已有相关类型或者细分功能类型两种，系统管理员可以根据实际需求自主选择确定相应的产品属性类型。
已有相关类型可以用于说明本次评论群创建请求所针对的软件产品对应的类型，例如理财类型、保险类型或者其他类型。一般来说，在软件产品存在相同类型(如产险A与产险B)或者相似类型(如产险A与寿险C)的已有软件产品，选取已有相关类型作为评论群创建请求的产品属性类型。主要原因在于，这些已有软件产品积累了相当数量的已有用户的情况下，可根据已有用户的用户行为数据确定待招募评论员，使得待招募评论员的确定更具有针对性。
细分功能类型是用于说明本次评论群创建请求所针对的软件产品所具体的功能。一般来说，在软件产品不存在相同类型或者相似类型的已有产品的情况下，如为新产品的前瞻性研究，市面上尚无完全对标的竞品及已有产品时，选取细分功能类型作为评论群创建请求的产品属性类型。具体地，通过对特定软件产品进行产品功能拆解，以获取该特定软件产品具体包括哪些细分功能(如登录功能、支付功能、拍照功能或者其他功能)，再寻找 与该细分功能相对应的已有用户，以便根据已有用户的用户行为数据确定待招募评论员。
S402：基于即时通讯工具，创建与产品标识相对应的目标评论群。
其中，即时通讯工具是设置在服务器上的可创建即时通信群的接口，该即时通讯工具包括但不限于微信接口、网易云信或者其他可实现即时通信功能的接口。具体地，服务器基于预先设置的即时通讯工具，创建与产品标识相对应的可实现即时通信功能的目标评论群，即该目标评论群主要针对产品标识对应的软件产品进行评论的通信群，使得该目标评论群的创建目的具有针对性。
S403：基于产品属性类型查询用户行为数据库，将预设时间段内与产品属性类型相匹配的活跃用户确定为待招募评论员。
其中，用户行为数据库是用于记录已有用户在相关软件产品的用户行为数据的数据库。用户行为数据是用于记录已有用户在相关软件产品的登录时间、登录时长和登录功能等数据。活跃用户是指根据已有用户的登录时间、登录时长和登录功能等用户行为数据确定达到活跃指标的用户。预设时间段是预先设置的用于限定所要采集的用户行为数据的时间，以避免将较长时间之前(如1年之前)活跃而最近一段时间(如3个月)不活跃的用户确定为待招募评论员，从而有助保障基于待招募评论员确定的目标评论员参与评论活动的积极性。待招募评论员是指根据已有用户参与活动的活跃程度确定的有较大可能被招募为最终的目标评论员的已有用户。
具体地，服务器基于产品属性类型查询对应的用户行为数据库，获取与产品属性类型相匹配的所有已有用户的用户行为数据；再筛选出登录时间在系统当前时间对应的预设时间段内(如近三个月)内的用户行为数据；最后，基于筛选出的用户行为数据分析该已有用户是否为活跃用户，以将该活跃用户确定为待招募评论员。其中，已有用户是否为活跃用户是指其用户行为数据分析是否达到相应的活跃指标确定。本实施例中，待招募评论员的确定，有助于保障基于待招募评论员确定的目标评论员参与评论活动的积极性。
S404：调用语音外呼平台对待招募评论员进行语音访问，获取待招募评论员对应的语音回访结果。
其中，语音外呼平台是用于实现语音外呼，以便进行语音访谈的平台。一般来说，在确定某一已有用户为待招募评论员之后，可以获取该已有用户的手机号码，通过调用语音外呼平台与待招募评论员用户进行1对1电话访谈，以确定其是否满足评论员招募条件，并征询其是否有参与评论的评论意愿，以便确定最终的目标评论员。语音回访结果是根据语音回访之后，确定某一待招募评论员是否满足评论员招募条件且是否有参与评论的评论意愿的结果。其中，语音回访结果包括符合评论员招募条件且有评论意愿、符合评论员招募条件但没有评论意愿、不符合评论员扫描条件但有评论意愿、不符合评论员招募条件且没有评论意愿等情况。
S405：若语音回访结果为符合评论员招募条件且有评论意愿，则将待招募评论员确定为目标评论员，邀请目标评论员加入目标评论群。
具体地，服务器在根据每一待招募评论员的语音回访结果确定该待招募评论员符合评论员招募条件且有评论意愿时，将其确定为目标评论员并邀请加入到目标评论群中，以使目标评论群中的每一目标评论员均为符合评论员招募条件且有评论意愿的评论员，使得目标评论员的确定经过特定的筛选，有助于保障后续针对软件产品进行评论时，目标评论员具有固定性，且其评论具有针对性，且目标评论员在目标评论群上发表的评论也具有即时性，有助于保障软件产品优化的效率，缩短优化周期。
本实施例所提供的软件产品测评处理方法中，基于即时通讯工具创建与产品标识相对应的目标评论群，使得该目标评论群的创建具有针对性，并保障后续通过目标评论群上获取的目标评论数据具有即时性。将预设时间段内与产品属性类型相匹配的活跃用户确定为待招募评论员，有助于保障基于待招募评论员确定的目标评论员参与评论活动的积极性。 根据待招募评论员的语音回访结果确定目标评论员，使得目标评论员的确定经过特定的筛选，有助于保障后续针对软件产品进行评论时，目标评论员具有固定性，且其评论具有针对性，且目标评论员在目标评论群上发表的评论也具有即时性，有助于保障软件产品优化的效率，缩短优化周期。
如图5所示，步骤S403，即基于产品属性类型查询用户行为数据库，将预设时间段内与产品属性类型相匹配的活跃用户确定为待招募评论员，包括：
S501：基于产品属性类型查询用户行为数据库，获取系统当前时间之前预设时间段内的用户行为数据确定为待分析用户数据，每一待分析用户数据包括一用户标识和登录时长。
其中，待分析用户数据是满足特定条件的可用于分析确定待招募评论员的数据。具体地，服务器基于产品属性类型查询对应的用户行为数据库，获取与产品属性类型相匹配的所有已有用户的用户行为数据；再筛选出登录时间在系统当前时间对应的预设时间段内(如近三个月)内的用户行为数据确定为待分析用户数据。该用户标识是用于唯一识别已有用户在某一软件产品中身份的标识，如手机号码。登录时长是用于登录某一软件产品过程中的时间间隔，可以理解为登录时间与退出时间之间的时间间隔。
S502：将登录时长大于预设时长的待分析用户数据，确定为有效行为数据。
其中，有效行为数据是指可以作为有效分析确定目标评论员对应的行为数据。本实施例中，将登录时长大于预设时长的待分析用户数据确定为有效行为数据，可以有效避免用户误点击所触发的登录而对用户行为数据的分析造成干扰，从而保障后续数据分析的准确性和效率。
S503：基于同一用户标识对应的所有有效行为数据，确定用户标识对应的目标评估指标。
其中，目标评估指标是最终确定是可以用于评估已有用户是否为活跃用户的指标。该目标评估指标包括但不限于登录次数、交易次数和常用功能信息等。具体地，服务器基于同一用户标识对应的所有有效行为数据，统计分析所有有效行为数据对应的登录次数、交易次数和常用功能信息等目标评估指标，以便判断已有用户是否为活跃用户。
S504：若目标评估指标符合对应的指标评估阈值，则将用户标识对应的用户确定为待招募评论员。
其中，指标评估阈值是预先设置的用于评估是否达到活跃用户标准的阈值。该指标评估阈值包括但不限于用于评估登录次数是否达到活跃用户标准的第一次数阈值、用于评估交易次数是否达到活跃用户标准的第二次数阈值和用于评估常用功能信息是否达到活跃用户标准的细分功能信息。本实施例中，可根据实际需求，将达到上述指标评估阈值的至少一个的用户确定为待招募评论员，以保证待招募评论员的活跃程度。
本实施例所提供的软件产品测评处理方法中，待分析用户数据是与产品属性类型相匹配且登录时间在系统当前时间之前预设时间段内的用户行为数据，可确保最终确定的待招募评论员是近期活跃用户；根据登录时长确定有效行为数据，有助于保障确定的待招募评论员的准确性和效率；根据有效行为数据确定的目标评估指标和指标评估阈值，确定待招募评论员，以保证待招募评论员的活跃程度，可有助于后续评论的积极性。
如图6所示，步骤S404，即调用语音外呼平台对待招募评论员进行语音访问，获取待招募评论员对应的语音回访结果，包括：
S601：获取与产品属性类型相对应的标准访谈话术，将标准访谈话术与待招募评论员对应的用户标识发送给语音外呼平台，获取语音外呼平台返回的回访录音数据。
其中，标准访谈话术是预先设置的用于引导待招募评论员回复是否满足评论员招募条件且是否有参与评论的评论意愿的话术。
本实施例中，服务器获取与产品属性类型相对应的标准访谈话术，将该标准访谈话术 与待招募评论员对应的用户标识(如手机号码或者可唯一确定对应的手机号码)发送给语音外呼平台，以便基于该语音外呼平台与用户标识对应的待招募评论员对应的移动终端进行电话通信，获取语音外呼平台返回的与该标准访谈话术相对应的回访录音数据。该回访录音数据是待招募评论员对标准访谈话术的语音回复。
S602：采用语音识别模型对回访录音数据进行识别，获取回访文本数据。
其中，语音识别模型是预先训练好的用于识别语音数据中的文本内容的模型。本实施例中的语音识别模型可采用语音静态解码网络，由于静态解码网络已经把搜索空间全部展开，因此其在进行文本翻译时，解码速度非常快，从而可快速获取回访文本数据。该语音静态解码网络是采用特定领域的训练语音数据进行训练所获取的静态解码网络，该特定领域的训练语音数据可以理解为预先存储的针对标准访谈话术进行回复的的语音数据。由于语音静态解码网络是基于特定领域的训练语音数据进行训练所获取的静态解码网络，使得其在对特定领域的回访录音数据进行识别时针对性强，使得解码准确率较高。
S603：对回访文本数据进行关键词提取，获取回访关键词。
具体地，服务器采用关键词提取算法对回访文本数据进行关键词提取，获取回访关键词。关键词提取算法是指从文本数据中提取其中的关键词的算法。例如，服务器先采用分词工具对回访文本数据进行分词，再采用去停用词算法对分词结果进行去停用词处理，以获取回访关键词。
S604：基于回访关键词查询评估结果对照表，获取待招募评论员对应的语音回访结果。
其中，评估结果对照表是预先设置用于评估是否满足评论员招募条件和是否有评论意愿的对照表。可以理解地，该评估结果对照表与标准访谈话术相匹配，即评估结果对照表可根据标准访谈话术中的引导问题确定不同结果相对应的预设关键词，以便根据该回访关键词与预设关键词的匹配结果，确定对应的语音回访结果。例如，标准访谈话术是设置“你是否意愿参与对软件产品A的项目优化项目”这一引导问题；在评估结果对照表中设置这一引导问题对应的预设关键词为“可以”、“愿意”和“想”等；则从这一引导问题对应的回访文本数据提取的回访关键词为“愿意”，则可以认定其有评论意愿。
本实施例所提供的软件产品测评处理方法中，通过对待招募评论员进行语音回访，获取回访录音数据，通过对该回访录音数据进行语音识别和关键词提取，以确定回访关键词之后，可查询评估结果对照表，从而快速获取待招募评论员对应的语音回访结果，以便实现对待招募评论员的进一步筛选，以使最终筛选出的目标评论员更符合软件产品的评论需求。
应理解，上述实施例中各步骤的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
在一实施例中，提供一种软件产品测评处理装置，该软件产品测评处理装置与上述实施例中软件产品测评处理方法一一对应。如图7所示，该软件产品测评处理装置包括目标评论数据获取模块701、目标关键词获取模块702、问题关键词获取模块703、出现频次获取模块704、高频问题确定模块705、目标优先级获取模块706、响应处理模块707、有效评论数据确定模块708、评论倾向结果获取模块709和目标评论结果获取模块710。各功能模块详细说明如下：目标评论数据获取模块701，用于基于即时通讯工具创建安装在产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识。目标关键词获取模块702，用于若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词。问题关键词获取模块703，用于根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词。出现频次获取模块704，用于对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词 对应的出现频次。高频问题确定模块705，用于若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题。目标优先级获取模块706，用于获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级。响应处理模块707，用于基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
优选地，在目标评论数据获取模块701之后，软件产品测评处理装置还包括：有效评论数据确定模块708，用于若所述目标评论数据携带所述任务标识，则获取所述任务标识对应的任务评论期限，在所述评论时间在所述任务评论期限内时，将所述目标评论数据确定为有效评论数据。评论倾向结果获取模块709，用于对所述有效评论数据进行分析，获取所述有效评论数据对应的评论问题和与所述评论问题相对应的评论倾向结果。目标评论结果获取模块710，用于统计与所述任务标识相关联的同一所述评论问题对应的每一所述评论倾向结果对应的评论数量，将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果。
优选地，在目标评论结果获取模块710之后，软件产品测评处理装置还包括：历史评论数据获取单元、评论参与比例获取单元、评论可用比例获取单元和评论奖励信息获取单元。历史评论数据获取单元，用于在系统当前时间为定时分析时间时，基于目标评论员的用户帐号查询评论数据库，获取评论时间在定时分析时间对应的数据统计周期内的与用户帐号相对应的历史评论数据。评论参与比例获取单元，用于基于历史评论数据查询与数据统计周期相对应的评论项目列表，获取历史评论数据对应的评论参与比例。评论可用比例获取单元，用于获取历史评论数据对应的历史倾向结果和目标评论结果，获取历史评论数据对应的评论可用比例。评论奖励信息获取单元，用于基于历史评论数据对应的评论参与比例和评论可用比例，获取用户帐号对应的评论奖励信息。
优选地，问题关键词获取模块703，包括：词性标注单元、待分析关键词确定单元、第一问题关键词确定单元和第二问题关键词确定单元。词性标注单元，用于采用词性标注工具对所述目标评论数据对应的目标关键词进行词性标注，获取每一所述目标关键词对应的词性。待分析关键词确定单元，用于将词性为预设词性的目标关键词确定为待分析关键词。匹配结果获取单元，用于采用匹配算法对所述待分析关键词和问题关键词库中的每一预设关键词进行逐一匹配处理，获取匹配结果。第一问题关键词确定单元，用于若存在所述匹配结果为匹配成功的预设关键词，则将对应的待分析关键词确定为所述目标评论数据对应的问题关键词。第二问题关键词确定单元，用于若不存在所述匹配结果为匹配成功的预设关键词，则根据所述目标评论数据的评论时间查询数据库，获取所述目标评论数据的评论时间之前的包含问题关键词的最近一条历史评论数据，根据最近一条所述历史评论数据的问题关键词确定所述目标评论数据对应的问题关键词。
优选地，目标评论数据获取模块701，包括：评论群创建请求获取单元、目标评论群创建单元、待招募评论员确定单元、语音回访结果获取单元和目标评论员确定单元。评论群创建请求获取单元，用于获取评论群创建请求，评论群创建请求包括产品标识和产品属性类型。目标评论群创建单元，用于基于即时通讯工具，创建与产品标识相对应的目标评论群。待招募评论员确定单元，用于基于产品属性类型查询用户行为数据库，将预设时间段内与产品属性类型相匹配的活跃用户确定为待招募评论员。语音回访结果获取单元，用于调用语音外呼平台对待招募评论员进行语音访问，获取待招募评论员对应的语音回访结果。目标评论员确定单元，用于若语音回访结果为符合评论员招募条件且有评论意愿，则将待招募评论员确定为目标评论员，邀请目标评论员加入目标评论群。
优选地，待招募评论员确定单元，包括：待分析用户数据获取子单元、有效行为数据确定子单元、目标评论指标确定子单元和待招募评论员确定子单元。待分析用户数据获取 子单元，用于基于产品属性类型查询用户行为数据库，获取系统当前时间之前预设时间段内的用户行为数据确定为待分析用户数据，每一待分析用户数据包括一用户标识和登录时长。有效行为数据确定子单元，用于将登录时长大于预设时长的待分析用户数据，确定为有效行为数据。目标评论指标确定子单元，用于基于同一用户标识对应的所有有效行为数据，确定用户标识对应的目标评估指标。待招募评论员确定子单元，用于若目标评估指标符合对应的指标评估阈值，则将用户标识对应的用户确定为待招募评论员。
优选地，语音回访结果获取单元，包括：回访录音数据获取子单元、回访文本数据获取子单元、回访关键词获取子单元和回访结果获取子单元。回访录音数据获取子单元，用于获取与产品属性类型相对应的标准访谈话术，将标准访谈话术与待招募评论员对应的用户标识发送给语音外呼平台，获取语音外呼平台返回的回访录音数据。回访文本数据获取子单元，用于采用语音识别模型对回访录音数据进行识别，获取回访文本数据。回访关键词获取子单元，用于对回访文本数据进行关键词提取，获取回访关键词。回访结果获取子单元，用于基于回访关键词查询评估结果对照表，获取待招募评论员对应的语音回访结果。
关于软件产品测评处理装置的具体限定可以参见上文中对于软件产品测评处理方法的限定，在此不再赘述。上述软件产品测评处理装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中，提供了一种计算机设备，该计算机设备可以是服务器，其内部结构图可以如图8所示。该计算机设备包括通过系统总线连接的处理器、存储器、网络接口和数据库。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统、计算机程序和数据库。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的数据库用于存储执行软件产品测评处理方法过程采集获取生成的数据，如目标评论结果。该计算机设备的网络接口用于与外部的终端通过网络连接通信。该计算机程序被处理器执行时以实现一种软件产品测评处理方法。
在一个实施例中，提供了一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，处理器执行计算机程序时实现上述实施例中软件产品测评处理方法，例如图2所示S201-S210，或者图3至图6中所示。或者，处理器执行计算机程序时实现软件产品测评处理装置这一实施例中的各模块/单元的功能，例如图7所示的各模块的功能。
在一实施例中，提供一计算机可读存储介质，该计算机可读存储介质上存储有计算机程序，该计算机程序被处理器执行时实现上述实施例中软件产品测评处理方法，例如图2所示S201-S210，或者图3至图6中所示。
所述的计算机程序可存储于一易失性或者非易失性计算机可读取存储介质中，该计算机程序在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限，RAM以多种形式可得，诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRAM)、存储器总线(Rambus)直接RAM(RDRAM)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。

Claims (20)

  1. 一种软件产品测评处理方法，其中，包括：
    基于即时通讯工具创建安装产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识；
    若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词；
    根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词；
    对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词对应的出现频次；
    若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题；
    获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级；
    基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
  2. 如权利要求1所述的软件产品测评处理方法，其中，在所述判断所述目标评论数据是否携带任务标识之后，所述软件产品测评处理方法还包括：
    若所述目标评论数据携带所述任务标识，则获取所述任务标识对应的任务评论期限，在所述评论时间在所述任务评论期限内时，将所述目标评论数据确定为有效评论数据；
    对所述有效评论数据进行分析，获取所述有效评论数据对应的评论问题和与所述评论问题相对应的评论倾向结果；
    统计与所述任务标识相关联的同一所述评论问题对应的每一所述评论倾向结果对应的评论数量，将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果。
  3. 如权利要求2所述的软件产品测评处理方法，其中，在所述将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果之后，所述软件产品测评处理方法还包括：
    在系统当前时间为定时分析时间时，基于所述目标评论员的用户帐号查询评论数据库，获取评论时间在所述定时分析时间对应的数据统计周期内的与所述用户帐号相对应的历史评论数据；
    基于所述历史评论数据查询与所述数据统计周期相对应的评论项目列表，获取所述历史评论数据对应的评论参与比例；
    获取所述历史评论数据对应的历史倾向结果和目标评论结果，获取所述历史评论数据对应的评论可用比例；
    基于所述历史评论数据对应的评论参与比例和评论可用比例，获取所述用户帐号对应的评论奖励信息。
  4. 如权利要求1所述的软件产品测评处理方法，其中，所述根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词，包括：
    采用词性标注工具对所述目标评论数据对应的目标关键词进行词性标注，获取每一所述目标关键词对应的词性；
    将词性为预设词性的目标关键词确定为待分析关键词；
    采用匹配算法对所述待分析关键词和问题关键词库中的每一预设关键词进行逐一匹 配处理，获取匹配结果；
    若存在所述匹配结果为匹配成功的预设关键词，则将对应的待分析关键词确定为所述目标评论数据对应的问题关键词；
    若不存在所述匹配结果为匹配成功的预设关键词，则根据所述目标评论数据的评论时间查询数据库，获取所述目标评论数据的评论时间之前的包含问题关键词的最近一条历史评论数据，根据最近一条所述历史评论数据的问题关键词确定所述目标评论数据对应的问题关键词。
  5. 如权利要求1所述的软件产品测评处理方法，其中，所述基于即时通讯工具创建安装所述产品测评平台上的任一特定软件产品对应的目标评论群，包括：
    获取评论群创建请求，所述评论群创建请求包括产品标识和产品属性类型；
    基于即时通讯工具，创建与所述产品标识相对应的目标评论群；
    基于所述产品属性类型查询用户行为数据库，将预设时间段内与所述产品属性类型相匹配的活跃用户确定为待招募评论员；
    调用语音外呼平台对所述待招募评论员进行语音访问，获取所述待招募评论员对应的语音回访结果；
    若所述语音回访结果为符合评论员招募条件且有评论意愿，则将所述待招募评论员确定为目标评论员，邀请所述目标评论员加入所述目标评论群。
  6. 如权利要求5所述的软件产品测评处理方法，其中，所述基于所述产品属性类型查询用户行为数据库，将预设时间段内与所述产品属性类型相匹配的活跃用户确定为待招募评论员，包括：
    基于所述产品属性类型查询用户行为数据库，获取系统当前时间之前预设时间段内的用户行为数据确定为待分析用户数据，每一所述待分析用户数据包括一用户标识和登录时长；
    将所述登录时长大于预设时长的待分析用户数据，确定为有效行为数据；
    基于同一所述用户标识对应的所有所述有效行为数据，确定所述用户标识对应的目标评估指标；
    若所述目标评估指标符合对应的指标评估阈值，则将所述用户标识对应的用户确定为待招募评论员。
  7. 如权利要求5所述的软件产品测评处理方法，其中，所述调用语音外呼平台对所述待招募评论员进行语音访问，获取所述待招募评论员对应的语音回访结果，包括：
    获取与所述产品属性类型相对应的标准访谈话术，将所述标准访谈话术与所述待招募评论员对应的用户标识发送给语音外呼平台，获取所述语音外呼平台返回的回访录音数据；
    采用语音识别模型对所述回访录音数据进行识别，获取回访文本数据；
    对所述回访文本数据进行关键词提取，获取回访关键词；
    基于所述回访关键词查询评估结果对照表，获取所述待招募评论员对应的语音回访结果。
  8. 一种软件产品测评处理装置，其中，包括：
    目标评论数据获取模块，用于基于即时通讯工具创建安装在产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识；
    目标关键词获取模块，用于若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词；
    问题关键词获取模块，用于根据所述目标评论数据对应的所述目标关键词和所述评论 时间，确定所述目标评论数据对应的问题关键词；
    出现频次获取模块，用于对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词对应的出现频次；
    高频问题确定模块，用于若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题；
    目标优先级获取模块，用于获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级；
    响应处理模块，用于基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
  9. 一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，其中，所述处理器执行所述计算机程序时实现如下步骤：
    基于即时通讯工具创建安装产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识；
    若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词；
    根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词；
    对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词对应的出现频次；
    若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题；
    获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级；
    基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
  10. 如权利要求9所述的计算机设备，其中，在所述判断所述目标评论数据是否携带任务标识之后，所述软件产品测评处理方法还包括：
    若所述目标评论数据携带所述任务标识，则获取所述任务标识对应的任务评论期限，在所述评论时间在所述任务评论期限内时，将所述目标评论数据确定为有效评论数据；
    对所述有效评论数据进行分析，获取所述有效评论数据对应的评论问题和与所述评论问题相对应的评论倾向结果；
    统计与所述任务标识相关联的同一所述评论问题对应的每一所述评论倾向结果对应的评论数量，将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果。
  11. 如权利要求10所述的计算机设备，其中，在所述将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果之后，所述软件产品测评处理方法还包括：
    在系统当前时间为定时分析时间时，基于所述目标评论员的用户帐号查询评论数据库，获取评论时间在所述定时分析时间对应的数据统计周期内的与所述用户帐号相对应的历史评论数据；
    基于所述历史评论数据查询与所述数据统计周期相对应的评论项目列表，获取所述历史评论数据对应的评论参与比例；
    获取所述历史评论数据对应的历史倾向结果和目标评论结果，获取所述历史评论数据 对应的评论可用比例；
    基于所述历史评论数据对应的评论参与比例和评论可用比例，获取所述用户帐号对应的评论奖励信息。
  12. 如权利要求9所述的计算机设备，其中，所述根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词，包括：
    采用词性标注工具对所述目标评论数据对应的目标关键词进行词性标注，获取每一所述目标关键词对应的词性；
    将词性为预设词性的目标关键词确定为待分析关键词；
    采用匹配算法对所述待分析关键词和问题关键词库中的每一预设关键词进行逐一匹配处理，获取匹配结果；
    若存在所述匹配结果为匹配成功的预设关键词，则将对应的待分析关键词确定为所述目标评论数据对应的问题关键词；
    若不存在所述匹配结果为匹配成功的预设关键词，则根据所述目标评论数据的评论时间查询数据库，获取所述目标评论数据的评论时间之前的包含问题关键词的最近一条历史评论数据，根据最近一条所述历史评论数据的问题关键词确定所述目标评论数据对应的问题关键词。
  13. 如权利要求9所述的计算机设备，其中，所述基于即时通讯工具创建安装所述产品测评平台上的任一特定软件产品对应的目标评论群，包括：
    获取评论群创建请求，所述评论群创建请求包括产品标识和产品属性类型；
    基于即时通讯工具，创建与所述产品标识相对应的目标评论群；
    基于所述产品属性类型查询用户行为数据库，将预设时间段内与所述产品属性类型相匹配的活跃用户确定为待招募评论员；
    调用语音外呼平台对所述待招募评论员进行语音访问，获取所述待招募评论员对应的语音回访结果；
    若所述语音回访结果为符合评论员招募条件且有评论意愿，则将所述待招募评论员确定为目标评论员，邀请所述目标评论员加入所述目标评论群。
  14. 如权利要求13所述的计算机设备，其中，所述基于所述产品属性类型查询用户行为数据库，将预设时间段内与所述产品属性类型相匹配的活跃用户确定为待招募评论员，包括：
    基于所述产品属性类型查询用户行为数据库，获取系统当前时间之前预设时间段内的用户行为数据确定为待分析用户数据，每一所述待分析用户数据包括一用户标识和登录时长；
    将所述登录时长大于预设时长的待分析用户数据，确定为有效行为数据；
    基于同一所述用户标识对应的所有所述有效行为数据，确定所述用户标识对应的目标评估指标；
    若所述目标评估指标符合对应的指标评估阈值，则将所述用户标识对应的用户确定为待招募评论员。
  15. 如权利要求13所述的计算机设备，其中，所述调用语音外呼平台对所述待招募评论员进行语音访问，获取所述待招募评论员对应的语音回访结果，包括：
    获取与所述产品属性类型相对应的标准访谈话术，将所述标准访谈话术与所述待招募评论员对应的用户标识发送给语音外呼平台，获取所述语音外呼平台返回的回访录音数据；
    采用语音识别模型对所述回访录音数据进行识别，获取回访文本数据；
    对所述回访文本数据进行关键词提取，获取回访关键词；
    基于所述回访关键词查询评估结果对照表，获取所述待招募评论员对应的语音回访结 果。
  16. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其中，所述计算机程序被处理器执行时实现如下步骤：
    基于即时通讯工具创建安装产品测评平台上的任一特定软件产品对应的目标评论群，获取所述目标评论群中任一目标评论员发表的目标评论数据和对应的评论时间，判断所述目标评论数据是否携带任务标识；
    若所述目标评论数据未携带所述任务标识，则采用关键词提取算法对所述目标评论数据进行关键词提取，获取所述目标评论数据对应的目标关键词；
    根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词；
    对系统当前时间对应的评论统计周期内的所述问题关键词进行频次统计，获取所述问题关键词对应的出现频次；
    若所述出现频次大于预设频次阈值，则将所述问题关键词确定为高频关键词，基于包含所述高频关键词的目标评论数据确定对应的高频问题；
    获取所述高频问题对应的问题发生频率和问题影响程度，依据所述问题发生频率和所述问题影响程度进行优先级分析，获取所述高频问题对应的目标优先级；
    基于所述目标优先级查询评论响应机制信息表，获取对应的评论响应机制，基于所述评论响应机制对所述高频问题进行响应处理。
  17. 如权利要求16所述的计算机可读存储介质，其中，在所述判断所述目标评论数据是否携带任务标识之后，所述软件产品测评处理方法还包括：
    若所述目标评论数据携带所述任务标识，则获取所述任务标识对应的任务评论期限，在所述评论时间在所述任务评论期限内时，将所述目标评论数据确定为有效评论数据；
    对所述有效评论数据进行分析，获取所述有效评论数据对应的评论问题和与所述评论问题相对应的评论倾向结果；
    统计与所述任务标识相关联的同一所述评论问题对应的每一所述评论倾向结果对应的评论数量，将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果。
  18. 如权利要求17所述的计算机可读存储介质，其中，在所述将评论数量最多的所述评论倾向结果确定为所述评论问题对应的目标评论结果之后，所述软件产品测评处理方法还包括：
    在系统当前时间为定时分析时间时，基于所述目标评论员的用户帐号查询评论数据库，获取评论时间在所述定时分析时间对应的数据统计周期内的与所述用户帐号相对应的历史评论数据；
    基于所述历史评论数据查询与所述数据统计周期相对应的评论项目列表，获取所述历史评论数据对应的评论参与比例；
    获取所述历史评论数据对应的历史倾向结果和目标评论结果，获取所述历史评论数据对应的评论可用比例；
    基于所述历史评论数据对应的评论参与比例和评论可用比例，获取所述用户帐号对应的评论奖励信息。
  19. 如权利要求16所述的计算机可读存储介质，其中，所述根据所述目标评论数据对应的所述目标关键词和所述评论时间，确定所述目标评论数据对应的问题关键词，包括：
    采用词性标注工具对所述目标评论数据对应的目标关键词进行词性标注，获取每一所述目标关键词对应的词性；
    将词性为预设词性的目标关键词确定为待分析关键词；
    采用匹配算法对所述待分析关键词和问题关键词库中的每一预设关键词进行逐一匹 配处理，获取匹配结果；
    若存在所述匹配结果为匹配成功的预设关键词，则将对应的待分析关键词确定为所述目标评论数据对应的问题关键词；
    若不存在所述匹配结果为匹配成功的预设关键词，则根据所述目标评论数据的评论时间查询数据库，获取所述目标评论数据的评论时间之前的包含问题关键词的最近一条历史评论数据，根据最近一条所述历史评论数据的问题关键词确定所述目标评论数据对应的问题关键词。
  20. 如权利要求16所述的计算机可读存储介质，其中，所述基于即时通讯工具创建安装所述产品测评平台上的任一特定软件产品对应的目标评论群，包括：
    获取评论群创建请求，所述评论群创建请求包括产品标识和产品属性类型；
    基于即时通讯工具，创建与所述产品标识相对应的目标评论群；
    基于所述产品属性类型查询用户行为数据库，将预设时间段内与所述产品属性类型相匹配的活跃用户确定为待招募评论员；
    调用语音外呼平台对所述待招募评论员进行语音访问，获取所述待招募评论员对应的语音回访结果；
    若所述语音回访结果为符合评论员招募条件且有评论意愿，则将所述待招募评论员确定为目标评论员，邀请所述目标评论员加入所述目标评论群。
PCT/CN2020/085579 2019-05-22 2020-04-20 软件产品测评处理方法、装置、计算机设备及存储介质 Ceased WO2020233309A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910430101.5A CN110263329B (zh) 2019-05-22 2019-05-22 软件产品测评处理方法、装置、计算机设备及存储介质
CN201910430101.5 2019-05-22

Publications (1)

Publication Number Publication Date
WO2020233309A1 true WO2020233309A1 (zh) 2020-11-26

Family

ID=67915096

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/085579 Ceased WO2020233309A1 (zh) 2019-05-22 2020-04-20 软件产品测评处理方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN110263329B (zh)
WO (1) WO2020233309A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112417300A (zh) * 2020-12-10 2021-02-26 平安普惠企业管理有限公司 产品漏洞方案查询方法、装置、电子设备及存储介质
CN113262467A (zh) * 2021-05-19 2021-08-17 北京小米移动软件有限公司 应用控制方法、装置及存储介质

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110263329B (zh) * 2019-05-22 2022-09-09 深圳壹账通智能科技有限公司 软件产品测评处理方法、装置、计算机设备及存储介质
CN111091473B (zh) * 2019-11-25 2023-08-15 泰康保险集团股份有限公司 保险问题分析处理方法和装置
CN111428159B (zh) * 2020-03-17 2024-07-23 建信金融科技有限责任公司 线上化分类方法和装置
CN114255064A (zh) * 2020-09-24 2022-03-29 苏州诗茉莉如汐文化传媒有限公司 一种cis企业形象设计服务管理系统
CN112508599B (zh) * 2020-11-13 2024-05-24 北京沃东天骏信息技术有限公司 信息反馈方法和装置
CN112529216A (zh) * 2020-12-02 2021-03-19 航天信息股份有限公司 一种一体化的运维方法及系统
CN113704568B (zh) * 2021-04-07 2025-08-08 腾讯科技（深圳）有限公司 内容识别方法、装置、设备及存储介质
CN113426132B (zh) * 2021-06-24 2023-09-05 咪咕互动娱乐有限公司 游戏优化方法、装置、设备及存储介质
CN114118937A (zh) * 2021-10-19 2022-03-01 北京百度网讯科技有限公司 基于任务的信息推荐方法、装置、电子设备及存储介质
CN114580981B (zh) * 2022-05-07 2022-08-02 广汽埃安新能源汽车有限公司 以用户需求驱动的项目调度方法、装置及电子设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110119243A1 (en) * 2009-10-30 2011-05-19 Evri Inc. Keyword-based search engine results using enhanced query strategies
CN103246598A (zh) * 2012-02-09 2013-08-14 广州博纳信息技术有限公司 软件测评远程测试方法
CN105468512A (zh) * 2014-09-05 2016-04-06 北京畅游天下网络技术有限公司 一种对软件质量进行评估的方法和系统
CN105787662A (zh) * 2016-02-25 2016-07-20 西北工业大学 基于属性的移动应用软件性能预测方法
CN108073703A (zh) * 2017-12-14 2018-05-25 郑州云海信息技术有限公司 一种评论信息获取方法、装置、设备及存储介质
CN110263329A (zh) * 2019-05-22 2019-09-20 深圳壹账通智能科技有限公司 软件产品测评处理方法、装置、计算机设备及存储介质

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2003276662A1 (en) * 2002-11-07 2004-06-07 Invoke Solutions, Inc. Survey system
CN104991956B (zh) * 2015-07-21 2018-07-31 中国人民解放军信息工程大学 基于主题概率模型的微博传播群体划分与账户活跃度评估方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110119243A1 (en) * 2009-10-30 2011-05-19 Evri Inc. Keyword-based search engine results using enhanced query strategies
CN103246598A (zh) * 2012-02-09 2013-08-14 广州博纳信息技术有限公司 软件测评远程测试方法
CN105468512A (zh) * 2014-09-05 2016-04-06 北京畅游天下网络技术有限公司 一种对软件质量进行评估的方法和系统
CN105787662A (zh) * 2016-02-25 2016-07-20 西北工业大学 基于属性的移动应用软件性能预测方法
CN108073703A (zh) * 2017-12-14 2018-05-25 郑州云海信息技术有限公司 一种评论信息获取方法、装置、设备及存储介质
CN110263329A (zh) * 2019-05-22 2019-09-20 深圳壹账通智能科技有限公司 软件产品测评处理方法、装置、计算机设备及存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112417300A (zh) * 2020-12-10 2021-02-26 平安普惠企业管理有限公司 产品漏洞方案查询方法、装置、电子设备及存储介质
CN113262467A (zh) * 2021-05-19 2021-08-17 北京小米移动软件有限公司 应用控制方法、装置及存储介质

Also Published As

Publication number Publication date
CN110263329A (zh) 2019-09-20
CN110263329B (zh) 2022-09-09

Similar Documents

Publication Publication Date Title
CN110263329B (zh) 软件产品测评处理方法、装置、计算机设备及存储介质
US10459954B1 (en) Dataset connector and crawler to identify data lineage and segment data
US10614077B2 (en) Computer system for automated assessment at scale of topic-specific social media impact
US9426289B2 (en) Techniques for topical customer service menu reconfiguration based on social media
US9646077B2 (en) Time-series analysis based on world event derived from unstructured content
US20190180195A1 (en) Systems and methods for training machine learning models using active learning
US10061845B2 (en) Analysis of unstructured computer text to generate themes and determine sentiment
CN117743371A (zh) 基于大语言模型的sql语句生成方法、装置、设备及介质
US11392835B2 (en) Employee concierge
US10847266B1 (en) Systems and methods for tracking goals
US20140189000A1 (en) Social media impact assessment
AU2020200806B2 (en) Virtual artificial intelligence based consultant
WO2019062078A1 (zh) 智能客服方法、电子装置及计算机可读存储介质
US20250156639A1 (en) Document analysis and management systems and methods
US10191985B1 (en) System and method for auto-curation of Q and A websites for search engine optimization
CN110494882A (zh) 决策支持系统及其相关方法
US20200409962A1 (en) Topic-specific reputation scoring and topic-specific endorsement notifications in a collaboration tool
US10147020B1 (en) System and method for computational disambiguation and prediction of dynamic hierarchical data structures
JP2022078008A (ja) 法令関連情報利用支援システム及びこれを用いた法令関連情報利用支援方法
Manda et al. Trade union membership and earnings in Kenyan manufacturing firms
US20100082361A1 (en) Apparatus, System and Method for Predicting Attitudinal Segments
US10803514B1 (en) Intelligent product plan dataset
US10803519B1 (en) Systems and methods for locating funding opportunities
CN116756419A (zh) 基于人工智能的信用卡权益推荐方法、装置、设备及介质
WO2019205381A1 (zh) 股票筛选方法、装置及计算机可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20808934

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20808934

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 22/03/2022)

122 Ep: pct application non-entry in european phase

Ref document number: 20808934

Country of ref document: EP

Kind code of ref document: A1