Statistical methods and systems for human resource management based on digitalization

Abstract

The present invention provides a human resource management statistics method and system based on digitalization, which receive a screening request from an evaluation end, obtain an application template of each candidate end and input it into a screening training model for question screening, obtain a customized question list corresponding to each candidate end, generate automatic question data corresponding to each candidate end according to the customized question list, obtain recorded video data generated by each candidate end based on the automatic question data, construct a backtracking node corresponding to each interaction segment in the recorded video data, generate an interactive backtracking chain corresponding to the candidate end according to each backtracking node, summarize the interactive backtracking chain of each candidate end, obtain the backtracking summary data and send it to the evaluation end.

Description

Technical Field

[0001] The present invention relates to data processing technology, and in particular to a human resource management statistical method and system based on digitization. Background Art

[0002] With the rapid development of economic globalization, human resource management is becoming more and more important to the development of enterprises. Interviews, as an important part of human resource management, cannot be ignored.

[0003] In the prior art, during the corporate interview process, manual interviews are usually conducted, specifically face-to-face interviews by interviewers. The entire recruitment process is time-consuming and the interview efficiency is low.

[0004] Therefore, how to conduct efficient interviews with candidates based on the interviewer's needs and improve interview efficiency has become an urgent problem that needs to be solved. Summary of the invention

[0005] The present invention provides a human resource management statistics method and system based on digitization, which can conduct efficient interviews with applicants in combination with the needs of the interviewers, thereby improving the interview efficiency.

[0006] The first aspect of the present invention provides a human resource management statistics method based on digitalization, comprising:

[0007] Receive the screening request from the evaluation end, obtain the application template of each candidate end and input it into the screening training model for question screening, and obtain a customized list of questions corresponding to each candidate end;

[0008] Generate automatic questioning data corresponding to each candidate terminal according to the customized question list, and obtain recorded video data generated by each candidate terminal based on the automatic questioning data;

[0009] Constructing a backtracking node corresponding to each interaction segment in the recorded video data, and generating an interaction backtracking chain corresponding to the candidate end according to each backtracking node;

[0010] The interactive backtracking chains of the candidate terminals are summarized and processed to obtain the backtracking summary data which is sent to the evaluation terminal.

[0011] Optionally, in a possible implementation of the first aspect, a screening request from an evaluation terminal is received, an application template of each candidate terminal is obtained, and inputted into a screening training model for question screening, and a customized list of questions corresponding to each candidate terminal is obtained, including:

[0012] According to the screening training model, the information type corresponding to each information column in the application template and the estimated duration configured by the evaluation end for each information type are obtained, and the information type of each information column in the application template is consistent;

[0013] Extracting the filled information in each of the information columns in the application template, traversing the question database based on the filled information and the estimated duration, and determining a plurality of selected questions in the question database corresponding to the corresponding information type;

[0014] Arrange the selected questions from large to small based on the screening weight of each selected question to obtain a list of sub-questions corresponding to the information type, and arrange the sub-question lists according to the preset arrangement order of each information type to obtain a customized list of questions corresponding to each candidate end.

[0015] Optionally, in a possible implementation of the first aspect, traversing a question database based on the filled-in information and the estimated duration to determine a plurality of selected questions in the question database corresponding to the corresponding information type includes: acquiring a screening level selected by the evaluation end, determining a question database corresponding to the screening level, acquiring a sub-database in the question database having the same information type as the filled-in information, wherein the question database includes a sub-database corresponding to each of the information types;

[0016] Acquire the screening time period selected by the evaluation end, determine the storage time of each historical question in the sub-database, and acquire the historical questions whose storage time is within the screening time period as one screening question;

[0017] Determine multiple information tags corresponding to the filled-in information according to the keyword entries corresponding to the filled-in information, traverse multiple custom tags corresponding to each of the first screening questions based on the information tags, and determine the custom tags identical to the information tags as the same tags;

[0018] Based on the ratio of the same number of the same labels in each of the first screening questions to the total number of labels, the information matching degree between the filled information and each of the first screening questions is obtained, and the first screening questions whose information matching degree is greater than the information matching threshold are obtained as the second screening questions;

[0019] According to the usage frequency and response time limit corresponding to each of the secondary screening questions, the screening weight of each of the secondary screening questions is determined, and based on the screening weight and the estimated time, the secondary screening questions are screened to obtain the selected questions corresponding to the information type.

[0020] Optionally, in a possible implementation of the first aspect, determining a screening weight of each of the secondary screening questions according to a usage frequency and a response time limit corresponding to each of the secondary screening questions, and screening the secondary screening questions based on the screening weight and the estimated time to obtain a selected question corresponding to the information type includes:

[0021] Obtaining the usage frequency of each of the secondary screening questions within the screening time period, and obtaining a type of adjustment coefficient according to the ratio of the usage frequency to the benchmark frequency;

[0022] Retrieving a time limit coefficient list corresponding to the information type, traversing the time limit coefficient list according to the preset response time limit corresponding to the secondary screening question, and determining that the preset coefficient corresponding to the time limit interval where the response time limit is located in the time limit coefficient list is the second type of adjustment coefficient;

[0023] A comprehensive adjustment coefficient is obtained according to the sum of the first adjustment coefficient and the second adjustment coefficient, and a screening weight corresponding to the secondary screening question is obtained by shifting the reference weight based on the comprehensive adjustment coefficient;

[0024] The secondary screening questions are arranged from large to small based on the screening weight to obtain a screening sequence, and the selected questions are determined for the screening sequence according to the estimated duration and the response time limit corresponding to each secondary screening question.

[0025] Optionally, in a possible implementation of the first aspect, determining the selected question for the screening sequence according to the estimated duration and the response time limit corresponding to each of the secondary screening questions includes:

[0026] Determine the first secondary screening question in the screening sequence as a candidate question; when the response time limit of the candidate question is less than or equal to the estimated time, determine the candidate question as a selected question; determine the secondary screening question adjacent to the selected question in the screening sequence as the next candidate question; and obtain the remaining time according to the time difference between the preset time and the response time limit of the selected question;

[0027] When the reply time limit of the question to be selected is greater than the estimated time, a variable time is obtained according to the time difference between the reply time limit of the question to be selected and the preset time. When the variable time is less than the time threshold, the question to be selected is determined to be the selected question, and the determination of the selected question is stopped.

[0028] When the change duration is greater than or equal to the duration threshold, the candidate question is deleted, and the first secondary screening question in the screening sequence is continuously determined as the next candidate question;

[0029] Repeat the above steps of determining the selected question, and use the remaining time obtained last time as the estimated time when determining the next selected question, and stop determining the selected question when the remaining time is less than or equal to the time threshold.

[0030] Optionally, in a possible implementation manner of the first aspect, generating automatic questioning data corresponding to each of the candidate terminals according to the customized question list, and acquiring recorded video data generated by each of the candidate terminals based on the automatic questioning data includes:

[0031] Obtaining question texts corresponding to each selected question in the question customization list, generating question voices according to the question texts, and generating automatic question videos corresponding to the selected questions based on the virtual digital human and the question voices;

[0032] Arrange the automatic questioning videos according to the arrangement order corresponding to the selected questions to obtain an automatic questioning sequence corresponding to the candidate terminal, wherein the automatic questioning data includes the automatic questioning sequence;

[0033] Switching to active communication mode based on the interview connection information of the candidate terminal, automatically displaying the automatic questioning video according to the automatic interaction area in the interaction interface, and displaying the active answering video of the candidate terminal based on the real-time interaction area in the interaction interface;

[0034] Switching to the passive communication mode according to the end information of the active communication mode, displaying the passive question video of the candidate terminal based on the real-time interaction area, and displaying the passive answer video corresponding to the question video based on the automatic interaction area;

[0035] The video data in the interactive interface in the active communication mode and the passive communication mode are recorded to obtain the recorded video data.

[0036] Optionally, in a possible implementation of the first aspect, switching to an active communication mode based on the interview connection information of the candidate terminal, automatically displaying the automatic questioning video according to an automatic interaction area in an interaction interface, and displaying the active answering video of the candidate terminal based on an instant interaction area in the interaction interface include:

[0037] Selecting the automatic questioning videos in the automatic questioning sequence as target questioning videos in sequence, displaying the target questioning videos based on the automatic interaction area, and obtaining the active answering videos of the candidate terminals in the real-time interaction area after the display is completed;

[0038] When the response time of the active answer video is equal to the response time limit of the target question video, or when a switching keyword is extracted based on the answer voice of the active answer video, the next target question video is displayed;

[0039] Switching to the passive communication mode according to the end information of the active communication mode, displaying the passive question video of the candidate terminal based on the real-time interaction area, and displaying the passive answer video corresponding to the question video based on the automatic interaction area, including:

[0040] When there is no next target question video, responding to the end information of the active communication mode and switching to the passive communication mode;

[0041] The passive question video of the candidate terminal in the real-time interaction area is obtained, and the corresponding passive answer video is retrieved according to the question keyword corresponding to the passive question video and displayed in the automatic interaction area.

[0042] Optionally, in a possible implementation manner of the first aspect, constructing a backtracking node corresponding to each interaction segment in the recorded video data, and generating an interaction backtracking chain corresponding to the candidate end according to each backtracking node, includes:

[0043] Acquire the video segments of the automatic questioning video and the active answering video corresponding to the recorded video data, and the video segments of the passive questioning video and the passive answering video corresponding to the recorded video data as interaction segments;

[0044] Constructing a recording timeline corresponding to the recorded video data, determining a recording segment corresponding to each of the interaction segments in the recording timeline, and constructing a backtracking node corresponding to each of the interaction segments at a center position of each of the recording segments;

[0045] Generate an overview image corresponding to each backtracking node according to the recorded frames and recorded voices corresponding to each of the interaction segments;

[0046] Retrieve a backtracking template, update a video display area in the backtracking template based on the interactive segment, update an overview area in the backtracking template based on the overview image to obtain backtracking data, bind the corresponding backtracking data and backtracking nodes to obtain the interactive backtracking chain.

[0047] Optionally, in a possible implementation manner of the first aspect, generating an overview image corresponding to each backtracking node according to the recorded frames and recorded voices corresponding to each of the interaction segments includes:

[0048] Obtaining a sub-recorded voice corresponding to the candidate end in the recorded voice, performing voice segmentation recognition on the sub-recorded voice to obtain voice segments, and extracting interactive texts corresponding to each of the voice segments;

[0049] Determine an intermediate moment corresponding to the speech segment, obtain a recording frame corresponding to the intermediate moment as a backtracking frame, and add the interactive text at a preset position of the backtracking frame to obtain a sub-overview image;

[0050] According to the time sequence and the preset arrangement sequence of each of the speech segments, each of the sub-overview images is spliced ​​to obtain an overview image corresponding to the corresponding backtracking node.

[0051] A second aspect of the present invention provides a human resource management statistics system based on digitalization, comprising:

[0052] A receiving module is used to receive a screening request from the evaluation end, obtain the application template of each candidate end and input it into the screening training model for question screening, and obtain a customized list of questions corresponding to each candidate end;

[0053] A generating module, configured to generate automatic questioning data corresponding to each candidate terminal according to the customized question list, and obtain recorded video data generated by each candidate terminal based on the automatic questioning data;

[0054] A construction module, used to construct a backtracking node corresponding to each interaction segment in the recorded video data, and generate an interaction backtracking chain corresponding to the candidate end according to each backtracking node;

[0055] The sending module is used to summarize the interactive backtracking chains of each candidate terminal, obtain the backtracking summary data and send it to the evaluation terminal.

[0056] The beneficial effects of the present invention are as follows:

[0057] 1. The present invention can combine the application information of the applicant and screen the historical interview questions in the enterprise's question database for multiple times, so as to generate a customized question list for the applicant, so that the interview questions in the interview process are more suitable for the applicant, and the screened interview questions meet the needs of the interviewer. After obtaining the customized question list, the present invention can combine the virtual digital human to generate automatic question data corresponding to the applicant, conduct an automatic interview with the applicant based on the automatic question data, and record the automatic interview process to generate recorded video data corresponding to the applicant. The interviewer can review the interview process of the applicant through the recorded video data, and can quickly evaluate the interview process of the applicant, thereby improving the efficiency of the interview.

[0058] 2. When screening historical interview questions in the enterprise's question database, the present invention first screens out primary screening questions that preliminarily meet the interviewer's needs in combination with the screening level and screening time period selected by the interviewer, and then screens the primary screening questions again in combination with the information filled in by the applicant in the application template, so as to obtain secondary screening questions that meet the applicant's application information, and then obtains the screening weights corresponding to the secondary screening questions in combination with the usage frequency and response time corresponding to the secondary screening questions, and screens the secondary screening questions again according to the screening weights and the estimated time preset by the interviewer, so as to obtain selected questions that meet the interviewer's needs and are highly matched with the applicant's information and meet the estimated time, and can improve the pertinence and effectiveness of the selected questions, and at the same time improve the interview efficiency.

[0059] 3. After the present invention obtains the recorded video data, it can combine the recorded video data to generate an interactive backtracking chain corresponding to the applicant, and can intuitively display the overall interview data of the applicant through the interactive backtracking chain, so that the interviewer can quickly view the interview process of the applicant, thereby improving the interview efficiency. When generating an interactive backtracking chain, the present invention can divide the recorded video data into multiple interactive segments, and generate an overview image in combination with the recorded frames and recorded voices corresponding to the interactive segments, so that the interviewer can quickly view the video data of the applicant's questions and answers during the interview, so that the interviewer can quickly preview the applicant's question and answer data, thereby improving the interview efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flowchart of a human resource management statistics method based on digitalization provided by an embodiment of the present invention;

[0061] Figure 2 is a schematic diagram of an interactive interface provided by an embodiment of the present invention;

[0062] Figure 3 is a schematic diagram of an interactive backtracking chain provided by an embodiment of the present invention;

[0063] Figure 4 It is a structural diagram of a digital human resource management statistics system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0066] See also Figure 1 , is a schematic diagram of a human resource management statistics method based on digitalization provided by an embodiment of the present invention, Figure 1 The execution subject of the method shown may be a software and / or hardware device. The execution subject of the present application may include but is not limited to at least one of the following: user equipment, network equipment, etc. Among them, the user equipment may include but is not limited to computers, smart phones, personal digital assistants (Personal Digital Assistant, referred to as: PDA) and the electronic devices mentioned above. The network equipment may include but is not limited to a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers, wherein cloud computing is a type of distributed computing, a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. It includes steps S1 to S4, as follows:

[0067] S1, receiving a screening request from the evaluation end, obtaining the application template of each candidate end and inputting it into a screening training model for question screening, and obtaining a customized list of questions corresponding to each candidate end.

[0068] This solution can screen interview questions based on the applicant's application information, so as to generate a customized question list for the applicant, that is, a customized question list. Subsequently, the list can be combined with a virtual digital person to generate automatic question data corresponding to the applicant, and the applicant can be automatically interviewed based on the automatic question data.

[0069] In actual applications, when applying for a job, applicants can enter application information such as basic personal information, work experience, etc. according to the application template pre-configured by the company. When the assessment end wants to interview the applicant, targeted questions can be customized based on the information filled in by the applicant in the template.

[0070] Specifically, the interviewer can send a request for screening interview questions based on the terminal. After receiving the screening request, the application template filled out by the applicant can be input into the screening training model, and the interview questions can be screened according to the application information filled out by the applicant in the application template. A customized list of questions corresponding to the applicant can be obtained. In some embodiments, the customized list of questions can be sent to the evaluation end. When the evaluation end wants to modify the question data in the customized list of questions, it can make corresponding modifications based on its own needs.

[0071] Among them, the evaluation end refers to the terminal held by the person in charge of the interview in the enterprise, for example, it can be a computer held by the interviewer. The screening request refers to the request for screening interview questions. The candidate end refers to the terminal held by the applicant. The application template refers to the template for applicants to fill in information when applying for an interview. The screening training model refers to a pre-trained model for customized screening of interview questions. The customized question list refers to a list of questions customized according to the information filled in by the applicant on the application template.

[0072] Based on the above embodiment, the specific implementation of step S1 may be:

[0073] S11, obtaining the information type corresponding to each information column in the application template and the estimated duration configured by the evaluation end for each information type according to the screening training model, and the information type of each information column in the application template is consistent.

[0074] Specifically, the application template contains multiple different information columns, such as an information column for filling in basic information and an information column for filling in work experience. The screening training model can identify the information type of each information column in the application template. Since the interview time of each applicant is limited, and the information filled in by the applicant may include multiple types, the interviewer may have different needs for understanding different types of information. For example, for personal information, the interviewer may only want to have a simple understanding, so during the interview, he may not spend a lot of time on asking questions about personal information. For work experience, the interviewer may want to have a deep understanding, so the interview time may be longer. Therefore, the interviewer can configure different interview durations for different information types based on his own needs. It is worth noting that the information types corresponding to each information column in the application template are consistent, but the content of the application information filled in by the applicant in the information column is inconsistent.

[0075] Among them, the information column refers to a data filling column used to fill in specific information, such as an area column for filling in work experience. The information type refers to the filling type corresponding to the information column, such as personal information, work experience, etc. The estimated duration refers to the total answer time of the interview questions corresponding to the information type configured by the interviewee based on his or her own needs.

[0076] S12, extracting the filled-in information in each information column in the application template, traversing the question database based on the filled-in information and the estimated duration, and determining a plurality of selected questions in the question database corresponding to the corresponding information type.

[0077] It is understandable that by recording interview questions from previous interviews, a database containing a large number of historical interview questions, i.e., a question database, can be formed. When screening interview questions for applicants, in order to obtain interview questions that are highly consistent with the applicant's information, the information filled in by the applicant in each information column in the application template can be extracted, and interview questions that are highly matched with the filled-in information can be obtained from the question database.

[0078] Specifically, since the filled-in information in each information column has a corresponding information type, and each information type has a corresponding number of historically stored interview questions in the question database, the question database can be traversed according to the filled-in information in each information column and the corresponding estimated duration, and the historical interview questions that highly match the filled-in information and meet the estimated duration can be determined as selected questions, thereby obtaining multiple selected questions corresponding to the corresponding information type in the question database. For example, when the information type is personal information, the selected questions that meet the requirements can be screened out in the question database based on the age, applied position, and other information filled in by the applicant in the information column corresponding to the personal information. Among them, the question database refers to a database that stores historical interview questions of different information types, and the selected questions refer to interview questions selected from the question database that correspond to the candidate end.

[0079] In some embodiments, the step S12 of "traversing the question database based on the filled-in information and the estimated duration, and determining a plurality of selected questions in the question database corresponding to the corresponding information type" includes the following steps: S121, obtaining the screening level selected by the evaluation end, determining the question database corresponding to the screening level, obtaining a sub-database in the question database with the same information type as the filled-in information, wherein the question database includes sub-databases corresponding to each of the information types.

[0080] The question database in this solution contains sub-databases corresponding to multiple information types, and the sub-database of each information type stores corresponding question data. Therefore, according to the information type corresponding to the filled-in information, the sub-database with the same information type as the filled-in information can be obtained from the question database, and the question database is pre-configured with corresponding screening levels. The historical storage data corresponding to different screening levels may be different. For example, the screening level A has the highest level, the screening level B has a lower level than the screening level A, and the screening level C has the lowest level. Then the screening level A can correspond to the historical storage data of the entire enterprise, the screening level B can correspond to the historical storage data of the department, and the screening level C can correspond to the historical storage data of the interviewee. The information types contained in the question databases corresponding to different screening levels are the same, but the stored question data are different. The higher the screening level, the more comprehensive the question data contained in the question database can be, and the larger the amount of data stored. For example, the question database corresponding to the screening level A can contain the historical question data of all employees of the enterprise, the question database corresponding to the screening level B can contain the historical question data of all employees of the department, and the question database corresponding to the screening level C can contain the historical question data of each employee.

[0081] Specifically, when determining the selected questions, the interviewer can first select the corresponding screening level according to the needs, and then can obtain the question database corresponding to the screening level. For example, when interviewer A wants to select interview questions from the historical question data of all employees in the company, then interviewer A can choose to select interview questions from the question database corresponding to screening level A. When interviewer B wants to select interview questions from the interviewer's personal historical question data, then interviewer B can choose to select interview questions from the question database corresponding to screening level C.

[0082] The screening level refers to a level that can be used to measure the amount of data stored in the problem database, and the sub-database refers to a database corresponding to each information type.

[0083] S122, obtaining the screening time period selected by the evaluation end, determining the storage time of each historical question in the sub-database, and obtaining the historical questions whose storage time is within the screening time period as a screening question.

[0084] In actual applications, the question data stored in the question database includes multiple question data of historical time. When screening questions, you can choose to screen the question data within the corresponding time period based on the needs of the interviewer. For example, the interviewer can choose the screening time period of the past month, so when screening, you can screen the question data within the past month. And each question data will record its corresponding storage time. When the same question is recorded multiple times, there will be multiple corresponding storage times, which can indicate that the historical question has been used multiple times. When the storage time of the historical question is within the screening time period, it can be considered that the historical question has been used within the screening time period. Therefore, the corresponding historical question can be determined as a screening question, and the screening question can be screened again later.

[0085] Among them, the screening time period refers to the time period that the interviewer chooses to screen the question data, for example, the screening time period can be within the past month, the storage time refers to the time for recording historical questions, and a screening question refers to a historical question whose storage time is within the screening time period.

[0086] S123, determining multiple information tags corresponding to the filled-in information according to the key words corresponding to the filled-in information, traversing multiple custom tags corresponding to each of the first screening questions based on the information tags, and determining the custom tags that are the same as the information tags as the same tags.

[0087] In actual applications, multiple key words can be extracted from the information entered by the applicant in the information column. For example, the key word extracted about age information can be 20 years old, and the key word extracted about job intention can be front-end development engineer, etc. The solution can pre-configure multiple information tags, and each information tag has a corresponding preset key word. According to the preset key words corresponding to the information tag, the key words that are the same as the preset key words of the information tag can be found from the multiple key words, thereby determining the multiple information tags corresponding to the filled-in information.

[0088] This solution will match each historical question with its corresponding custom label. For example, when the information type corresponding to the historical question is basic personal information, its corresponding custom label may be 25-30 years old, front-end development, etc. The label can indicate the applicable age group, matching position and other information corresponding to the historical question. Since each one-time screening question is configured with multiple corresponding custom labels, the custom labels corresponding to each one-time screening question can be traversed, and each custom label can be compared with the information label corresponding to the filled-in information to find out those custom labels that are exactly the same as the information label and determine them as the same label. The same label can be used to determine the degree of information overlap between the filled-in information and the one-time screening question.

[0089] Among them, the keyword entry refers to the entry extracted from the filled-in information. For example, when the age filled in the filled-in information is 20 years old, the extracted keyword entry can be 20 years old. The information tag refers to the label corresponding to the filled-in information of the applicant. Since the filled-in information may involve multiple aspects of information, such as age, gender, job position, etc., it may correspond to multiple information tags. Custom tags refer to information tags configured for a screening question.

[0090] S124, based on the ratio of the same number of the same labels in each of the first screening questions to the total number of labels, obtain the information matching degree between the filled-in information and each of the first screening questions, and obtain the first screening questions whose information matching degree is greater than the information matching threshold as the second screening questions.

[0091] Among them, the information matching degree refers to the numerical value reflecting the matching degree between the screening question and the filled-in information. The higher the information matching degree, the higher the matching degree between the filled-in information and the primary screening question. The information matching threshold refers to the minimum value used to judge whether the primary screening question is sufficiently matched with the filled-in information. When the information matching degree is greater than the information matching threshold, the primary screening question can be retained and determined as the secondary screening question. In order to obtain interview questions that meet the estimated duration, the secondary screening questions can be screened again later.

[0092] S125, determining a screening weight for each of the secondary screening questions according to the usage frequency and response time limit corresponding to each of the secondary screening questions, and screening the secondary screening questions based on the screening weight and the estimated time to obtain selected questions corresponding to the information type.

[0093] In actual applications, the frequency of use of secondary screening questions during the screening time period can be used to determine whether the secondary screening questions are frequently used during the screening time period. If the frequency of use of secondary screening questions is high, it can be considered that it is frequently used during the screening time period, and it may be a common or important question, which can be given priority. In order to screen out interview questions that meet the interviewee's estimated time, the secondary screening questions can be screened again in combination with the response time limit.

[0094] Based on this, the screening weight of the secondary screening questions can be determined according to the usage frequency and response time limit corresponding to the secondary screening questions. After obtaining the screening weight, the secondary screening questions can be screened again in combination with the screening weight and estimated duration, so that selected questions that meet the needs of the interviewers, highly match the information of the candidates and meet the estimated duration can be obtained.

[0095] Among them, usage frequency refers to the number of times the secondary screening questions are used within the screening time period, response time limit refers to the preset answer time for the secondary screening questions, and screening weight refers to the weight determined by usage frequency and response time limit, which can be used to measure the priority of each selected question.

[0096] Through the above implementation, the selected information obtained not only meets the needs of the interviewer, but also closely fits the application information of the applicant.

[0097] Based on the above embodiment, the specific implementation of step S125 may be:

[0098] S1251, obtaining the usage frequency of each of the secondary screening questions within the screening time period, and obtaining a type of adjustment coefficient according to the ratio of the usage frequency to the reference frequency.

[0099] Specifically, the usage frequency of the secondary screening questions in the screening time period can be accurately determined according to the number of times the secondary screening questions are stored in the screening time period. The base frequency refers to the preset standard usage frequency, and the first adjustment coefficient refers to the ratio of the usage frequency to the base frequency. The higher the usage frequency, the larger the first adjustment coefficient. The secondary screening questions with higher usage frequency can be screened preferentially based on the coefficient.

[0100] S1252, retrieve the time limit coefficient list corresponding to the information type, traverse the time limit coefficient list according to the preset response time limit corresponding to the secondary screening question, and determine that the preset coefficient corresponding to the time limit interval where the response time limit in the time limit coefficient list is located is the second type of adjustment coefficient.

[0101] Since the time requirements for different information types may be different, for example, personal information may require a simple and quick understanding. Therefore, when selecting such questions, questions with a shorter answer time may be given priority. Therefore, a list of time limit coefficients corresponding to different information types may be configured in advance. The list may be set in combination with the time requirements of the corresponding information types. Multiple time limit intervals may be set in advance, and each time limit interval may be configured with a corresponding preset coefficient based on the requirements. For example, for the time limit coefficient list corresponding to personal information, the shorter the time limit interval, the larger the corresponding preset coefficient may be, so that question data that meets the corresponding type of time requirements may be given priority.

[0102] Specifically, in order to determine the second type of adjustment coefficient corresponding to the secondary screening question, we can first call up the time limit coefficient list corresponding to the information type, and traverse and search in the time limit coefficient list according to the preset response time limit of the secondary screening question. It can be determined that the response time limit falls within which the preset time limit interval falls. After determining the time limit interval, the preset coefficient corresponding to the interval can be directly determined as the second type of adjustment coefficient of the current secondary screening question.

[0103] S1253, obtaining a comprehensive adjustment coefficient according to the sum of the first-category adjustment coefficient and the second-category adjustment coefficient, and offsetting the benchmark weight based on the comprehensive adjustment coefficient to obtain a screening weight corresponding to the secondary screening question.

[0104] By combining the comprehensive adjustment coefficient to offset the base weight and obtain the screening weight, you can prioritize the secondary screening questions that are used frequently and meet time requirements. Among them, the comprehensive adjustment coefficient refers to the number obtained by adding the first adjustment coefficient and the second adjustment coefficient, and the base weight refers to the preset weight.

[0105] S1254, arranging the secondary screening questions from large to small based on the screening weight to obtain a screening sequence, and determining the selected questions for the screening sequence according to the estimated duration and the response time limit corresponding to each secondary screening question.

[0106] In order to ensure that the final selected questions meet the estimated time, the secondary screening questions can be finally selected based on the response time limit and the estimated time.

[0107] The screening sequence refers to a sequence obtained by arranging each secondary screening question from large to small according to the screening weight.

[0108] Through the above implementation, selected questions that meet the needs of the interviewer and are highly matched with the information of the applicant can be obtained, so that a customized question customization list can be generated for the applicant.

[0109] In some embodiments, the step S1254 of "determining the selected questions for the screening sequence according to the estimated duration and the response time limits corresponding to the secondary screening questions" includes the following steps:

[0110] S12541, determine the first secondary screening question in the screening sequence as a candidate question; when the response time limit of the candidate question is less than or equal to the estimated time, determine the candidate question as a selected question; determine the secondary screening question adjacent to the selected question in the screening sequence as the next candidate question; and obtain the remaining time according to the time difference between the estimated time and the response time limit of the selected question.

[0111] In order to ensure that the total duration of the selected questions can meet the estimated duration, this solution will select secondary screening questions whose total response time is close to the estimated time as the final selected questions.

[0112] Specifically, when the response time limit for the current question to be selected is less than or equal to the estimated time, then it can be considered that the current question to be selected can be answered within the estimated time, and the response time limit for the question to be selected meets the estimated time, so it can be determined as the selected question, and the remaining time is determined as the basis for the subsequent acquisition of the next selected question. For example, when the estimated time is 10 minutes and the response time limit for the question to be selected is 6 minutes, the response time limit is less than the estimated time, so the question to be selected can be determined as the selected question, and the corresponding remaining time is 4 minutes, which can be used for subsequent judgment.

[0113] S12542, when the reply time limit of the question to be selected is greater than the estimated time, the variable time is obtained according to the time difference between the reply time limit of the question to be selected and the preset time. When the variable time is less than the time threshold, the question to be selected is determined to be the selected question, and the determination of the selected question is stopped.

[0114] When the response time limit of a candidate question exceeds the estimated time, if the excess time is small, it can be considered that the response time limit of the candidate question is still relatively close to the estimated time. At this time, the candidate question can be determined as the selected question.

[0115] Specifically, the time difference between the two, that is, the variable time, can be calculated. The variable time reflects the amount of time that exceeds the estimated time. By comparing the variable time and the time threshold, it can be determined whether the response time limit for the corresponding question to be selected is closer to the interviewee's estimated time. If the variable time is less than the preset time threshold, it means that the amount of time that exceeds the estimated time is very small. It can be considered that the response time limit for the question to be selected is closer to the estimated time. Therefore, this question to be selected can be determined as the selected question.

[0116] Since the total duration of the selected questions is already close to the estimated duration, the further determination of the selected questions can be stopped to avoid extending the interview time. For example, when the estimated duration is 10 minutes, the response time limit for the selected questions is 11 minutes, and the preset duration threshold is 2 minutes, the variable duration can be 1 minute. Since the variable duration is less than the duration threshold, the selected questions can be determined as the selected questions.

[0117] The variable duration refers to the additional duration that the response time exceeds the estimated duration, and the duration threshold refers to the maximum value of the duration that the response time is close to the estimated duration.

[0118] S12543, when the change duration is greater than or equal to the duration threshold, delete the candidate question, and continue to determine the first secondary screening question in the screening sequence as the next candidate question.

[0119] Specifically, when the change time is greater than the time threshold, it can be explained that the response time limit of the candidate question is significantly different from the estimated time, and the response time limit of the candidate question does not meet the interviewer's estimated time requirement. Therefore, the candidate question can be deleted. For example, when the estimated time is 10 minutes, the response time limit of the candidate question is 13 minutes, and the preset time threshold is 2 minutes, the change time can be 3 minutes. Since the change time is greater than the time threshold, the candidate question can be deleted.

[0120] After deleting the candidate question, the second secondary screening question in the original screening sequence will become the first secondary screening question, so the first secondary screening question can be selected from the screening sequence again as the next candidate question.

[0121] S12544, repeat the above steps of determining the selected question, and use the remaining time obtained last time as the estimated time when determining the next selected question, and stop determining the selected question when the remaining time is less than or equal to the time threshold.

[0122] Specifically, if the selected question has not been determined yet, that is, when the above-mentioned change time is greater than the time threshold, the determined question to be selected can continue to be compared with the estimated time until the first selected question is determined. If a selected question has been determined and the next selected question can be continued, that is, when the above-mentioned response time is less than or equal to the estimated time, the secondary screening question adjacent to the selected question in the screening sequence can be used as the next question to be selected, and then the response time of the question to be selected is compared with the remaining time. The selected question is continued to be determined through the comparison result, and the remaining time of the previous cycle is used for the next comparison. If the remaining time is less than or equal to the time threshold, it means that the total response time of the selected question at this time is also close to the estimated time, and the determination of the selected question can also be stopped. Through the above-mentioned cycle steps, the total time of the selected question can be made closer to the estimated time.

[0123] Through the above implementation, interview questions can be screened in combination with the needs of the interviewer, so that the final selected questions meet the needs of the interviewer.

[0124] S13, arranging each of the selected questions from large to small based on the screening weight of each of the selected questions to obtain a list of sub-questions corresponding to the information type, arranging each of the sub-question lists according to a preset arrangement order of each of the information types to obtain a customized list of questions corresponding to each of the candidate terminals.

[0125] It is understandable that since there are multiple selected questions, in order to ensure that the selected questions that meet time requirements and are used more frequently can be selected first, the multiple selected questions can be sorted to obtain a customized list of questions corresponding to the applicants.

[0126] Specifically, first, according to the screening weights corresponding to the obtained selected questions, the selected questions corresponding to each information type can be arranged in descending order according to the screening weights, and multiple sub-question lists corresponding to each information type can be obtained. Since each information type has a corresponding preset arrangement order, for example, during an interview, questions about personal information types may be asked first, and then questions about work experience types. Then, when sorting, the sub-question list of personal information types may be arranged first, and then the sub-question list of work experience types. Therefore, each sub-question list can be arranged according to the arrangement order of the information type to obtain a customized list of questions corresponding to the applicant.

[0127] The sub-question list refers to a table in which the selected questions corresponding to each information type are arranged, and the preset arrangement order refers to a preset order for arranging the information types.

[0128] Through the above implementation, it is possible to obtain selected questions that meet the needs of the interviewer and are more suitable for the applicant.

[0129] S2, generating automatic questioning data corresponding to each of the candidate terminals according to the customized question list, and obtaining recorded video data generated by each of the candidate terminals based on the automatic questioning data.

[0130] Specifically, after obtaining the customized list of questions, this solution can combine the virtual digital human to generate automatic question data corresponding to the applicant, and automatically interview the applicant based on the automatic question data. In addition, this solution can record the automatic interview process and generate recorded video data corresponding to each applicant, so that subsequent interviewers can quickly view the interview process of the applicant, thereby quickly analyzing and evaluating the interview process of the interviewer.

[0131] The automatic question data refers to the video data of questions asked during the automatic interview of the applicant.

[0132] In some embodiments, step S2 includes S21 to S25, which are specifically as follows:

[0133] S21, obtaining the question text corresponding to each selected question in the question customization list, generating a question voice according to the question text, and generating an automatic question video corresponding to the selected question based on the virtual digital human and the question voice.

[0134] In actual applications, for each selected question in the customized question list, the corresponding question text can be obtained. When obtaining the automatic question video, a virtual digital human model can be created through 3D modeling first. The model can display details such as expressions and movements, and the expressions and movements of the virtual digital human can be configured accordingly according to the needs of the interviewer. Then the question text of the selected question is input into the speech synthesis system, and the question text is converted into a natural and fluent voice, i.e., the question voice, by the speech synthesis system. The question voice can be configured accordingly according to the needs of the interviewer, for example, it can be set to a youth voice. After obtaining the question voice, the question voice data can be fused with the model data of the virtual digital human, so that the question voice can be combined with the movements and expressions of the virtual digital human to generate a video that can display the question data of the selected question, i.e., the automatic question video.

[0135] Among them, question text refers to the text data corresponding to the selected question, question voice refers to the voice corresponding to the question text, virtual digital human refers to a digitized virtual image with similar appearance and movement characteristics to humans, and automatic question video refers to a video that can display the question data of the selected question.

[0136] S22, arranging each of the automatic questioning videos according to the arrangement order corresponding to each of the selected questions, to obtain an automatic questioning sequence corresponding to the corresponding candidate terminal, wherein the automatic questioning data includes an automatic questioning sequence.

[0137] Specifically, each selected question has a corresponding arrangement order in the question customization list. The automatic questioning videos corresponding to each selected question are arranged in the arrangement order to obtain an automatic questioning sequence, and the automatic questioning videos can be automatically played according to the automatic questioning sequence.

[0138] S23, switching to active communication mode based on the interview connection information of the candidate terminal, automatically displaying the automatic questioning video according to the automatic interaction area in the interaction interface, and displaying the active answering video of the candidate terminal based on the real-time interaction area in the interaction interface.

[0139] See also Figure 2 , is a schematic diagram of an interactive interface provided by an embodiment of the present invention. The interactive interface is divided into two parts: an automatic interactive area and a real-time interactive area. The automatic interactive area can be used to display the video data of the interviewer during the interview process. For example, the automatic questioning video can be automatically displayed. Figure 2 The left area in the real-time interactive area can be used to display the video data of the applicant during the interview process, such as Figure 2 The right area in .

[0140] In some embodiments, the server can generate a corresponding interview link for the interview and send it to the candidate end. After the applicant clicks the interview link through the candidate end, the server can respond to the interview connection information and automatically switch to the active communication mode. In the active communication mode, the automatic interaction area in the interactive interface will display the automatic question video according to the automatic question sequence, and the applicant can answer after each automatic question video is displayed.

[0141] Among them, interview connection information refers to the connection information used to indicate joining the interview process, active communication mode refers to the communication mode of actively displaying automatic question videos and candidates answering questions based on the automatic question videos, interactive interface refers to the interface for displaying video data of the interview process, and active answer video refers to the video of the candidate answering selected questions.

[0142] Based on the above embodiment, the specific implementation of step S23 may be:

[0143] S231, sequentially selecting the automatic questioning videos in the automatic questioning sequence as target questioning videos, displaying the target questioning videos based on the automatic interaction area, and after the display is completed, obtaining the active answering videos of the candidate terminals in the real-time interaction area.

[0144] Specifically, according to the order of the automatic question sequence, the automatic question videos can be selected in sequence as the target question videos, and the target question videos can be displayed in the automatic interaction area. After each target question video is displayed, the applicant can answer the target question video, and after the target question video is displayed, the video data of the applicant in the instant interaction area can be timed, and the recorded video data can be used as the active answer video. The recorded time is the response time of the active answer video. Later, it can be determined whether to display the next target question video based on the response time of the active answer video. Among them, the target question video refers to the automatic question video selected from the automatic question sequence and displayed in the automatic interaction area each time.

[0145] S232, when the response time of the active answer video is equal to the response time limit corresponding to the target question video, or when a switching keyword is extracted based on the answer voice of the active answer video, the next target question video is displayed.

[0146] In actual applications, each selected question has a corresponding preset response time limit, so the target question video corresponding to each selected question also has the same preset response time limit. When the target question videos are selected in turn and displayed automatically, after each target question video is displayed, it is necessary to determine whether to display the next target question video.

[0147] Specifically, after each target question video is displayed, an answer time equal to the response time limit corresponding to the current target question video can be reserved for the applicant to answer the corresponding question. When the response time of the active answer video is equal to the response time limit, it means that the answer time reserved for the applicant has expired. Regardless of whether the applicant's answer is completed at this time, the video will automatically switch to the next target question video for playback.

[0148] If the applicant has answered the corresponding question in advance within the response time limit, the applicant's answer voice data can be extracted, the answer voice data can be converted into voice, and the corresponding text data can be obtained. When the corresponding switching keyword is extracted from the text data, for example, when the answer is completed from the corresponding text data, the next target question video can be switched. Among them, the response time refers to the length of time the video data of the applicant's answer is recorded, the answer voice refers to the voice of the applicant in the active answer video, and the switching keyword refers to the keyword that can represent the switching information.

[0149] S24, switching to the passive communication mode according to the end information of the active communication mode, displaying the passive question video of the candidate terminal based on the real-time interaction area, and displaying the passive answer video corresponding to the question video based on the automatic interaction area.

[0150] Among them, the passive communication mode refers to a communication mode of displaying a passive answer video based on the candidate's passive question video. When the active communication mode ends, it can be switched to the passive communication mode. The passive question video refers to a video of the candidate asking questions, and the passive answer video refers to a video answering the questions asked by the candidate.

[0151] Based on the above embodiment, the specific implementation of step S24 may be:

[0152] S241, when there is no next target question video, respond to the end information of the active communication mode and switch to the passive communication mode.

[0153] Specifically, when all the automatic question videos in the automatic question sequence have been played and the applicant has finished answering the corresponding questions, there is no next target question video at this time. It can be considered that the active communication mode has ended, and the end information can be responded to and automatically switched to the passive communication mode. In the passive communication mode, the applicant can ask questions.

[0154] S242, obtaining the passive question video of the candidate terminal in the real-time interaction area, and calling the corresponding passive answer video according to the question keyword corresponding to the passive question video to display it in the automatic interaction area.

[0155] Specifically, the applicant can ask questions that he or she wants to know about the interview, and when asking questions, the video data of the applicant in the real-time interaction area can be recorded, and the recorded video data within the corresponding length of time can be used as a passive question video. For each passive question video, the corresponding voice question data can be obtained, and the corresponding question keywords can be extracted from the voice question data. Then, based on the question keywords, the corresponding passive answer video can be retrieved from a pre-prepared database, and the passive answer video can be displayed in the automatic interaction area. For example, when the applicant asks about the company's working hours, the passive answer video related to the working hours can be retrieved and displayed in the automatic interaction area.

[0156] In addition, in some embodiments, if in the passive communication mode, there is no passive answer video corresponding to the passive question video, a preset follow-up video can be retrieved for display in the automatic interaction area. By playing back the video, the applicant can be informed that a follow-up visit will be conducted after querying the passive question questions, and the questions raised by the applicant can be recorded. Among them, the follow-up video refers to a video that can be used to inform the applicant that a follow-up visit will be conducted later.

[0157] S25, recording the video data in the interactive interface in the active communication mode and the passive communication mode to obtain the recorded video data.

[0158] In order to display the interview data of the applicant during the interview process to the interviewer for review, this solution can record the video data of the applicant during the interview process to obtain the recorded video data of the entire interview process.

[0159] Through the above implementation method, it can be ensured that every link of the interview is fully recorded, which is helpful for the subsequent quick review of the interview process and improves the interview efficiency.

[0160] S3, constructing a backtracking node corresponding to each interaction segment in the recorded video data, and generating an interaction backtracking chain corresponding to the candidate end according to each backtracking node.

[0161] It is understandable that after obtaining the recorded video data of each candidate's interview process, in order to enable the interviewer to quickly view the candidate's interview process and improve the interview efficiency, this solution can combine the recorded video data to generate an interactive backtracking chain corresponding to the candidate, and the overall interview data of the candidate can be intuitively displayed through the interactive backtracking chain.

[0162] Specifically, since the recorded video data may be long, in order to facilitate the interviewee to quickly view the video segment corresponding to each question and answer data, this solution can construct a node corresponding to each question and answer data, that is, a backtracking node, so that the interviewee can quickly browse the video segment corresponding to each question and answer data by clicking the backtracking node, and a corresponding interactive backtracking chain can be generated according to the backtracking node.

[0163] Among them, the interaction segment is composed of the video segment of the question and answer data corresponding to each question, the traceback node refers to the node that can display the video data corresponding to the interaction segment, and the interaction traceback chain refers to the traceback chain that can intuitively display each question and answer video segment in the recorded video data of the applicant.

[0164] In some embodiments, the interactive backtracking chain may be generated in the following manner:

[0165] S31, obtaining the video segments of the automatic questioning video and the active answering video corresponding to the recorded video data, and the video segments of the passive questioning video and the passive answering video corresponding to the recorded video data as interaction segments.

[0166] This solution will record the question time of each question and the applicant's answer time, so that the question time period and answer time period corresponding to each question can be obtained, and therefore the video data of each question video and each answer video in the corresponding time period can be obtained. For example, during the applicant's answer time period, the video data in the instant interaction area can be used as the video segment corresponding to the active answer video, and the video segment of each automatic question video and its corresponding video segment of the active answer video can together constitute an interaction segment, and the video segment of the passive question video and its corresponding video segment of the passive answer video can together constitute an interaction segment.

[0167] S32, constructing a recording timeline corresponding to the recorded video data, determining a recording segment corresponding to each of the interaction segments in the recording timeline, and constructing a backtracking node corresponding to each of the interaction segments at a center position of each of the recording segments.

[0168] Specifically, when displaying the recorded video data, a recording timeline corresponding to the recording time of the recorded video data can be constructed. In order to map the interaction segments to the recording timeline, the start and end times of each interaction segment in the recorded video data can be first identified, so that the time period corresponding to each interaction segment can be obtained, and the axis segments corresponding to each time period on the recording timeline are determined as the recording segments corresponding to each interaction segment in the recording timeline. In order to enable the interviewee to quickly locate the question and answer video segment corresponding to each question, a backtracking node corresponding to the interaction segment can be constructed at the center of the recording segment.

[0169] The recording timeline refers to an axis that can be used to display the time process of the entire recorded video, and the recording segment refers to the axis segment corresponding to the interaction segment on the recording timeline.

[0170] S33, generating an overview image corresponding to each backtracking node according to the recorded frames and recorded voices corresponding to each of the interaction segments.

[0171] It is understandable that in order to allow the interviewer to quickly review the candidate's question and answer data, an overview image can be generated by combining the recorded frames and recorded voice corresponding to the interaction segment, and the candidate's question and answer data can be quickly previewed through the overview image, thereby improving the efficiency of retrospective review.

[0172] Among them, the recorded frame refers to the video frame corresponding to the video data of the interaction segment, the recorded voice refers to the voice corresponding to the video data corresponding to the interaction segment, and the overview image refers to an image that can be used to preview the question and answer data of the applicant.

[0173] Based on the above embodiment, the specific implementation of step S33 may be:

[0174] S331, obtaining a sub-recorded voice corresponding to the candidate end in the recorded voice, performing voice segmentation recognition on the sub-recorded voice to obtain voice segments, and extracting interactive texts corresponding to each of the voice segments.

[0175] Specifically, the recorded voice includes the recorded voice of the virtual digital person and the recorded voice of the applicant. In order to quickly preview the question and answer situation of the applicant, the sub-recorded voice corresponding to the applicant can be identified from the recorded voice, and the sub-recorded voice can be segmented. For example, sentence recognition can be performed through voice feature analysis technology, and the continuous voice stream can be divided into voice segments with complete meanings, thereby obtaining multiple voice segments. By performing voice recognition on each voice segment, it can be converted into text form, that is, interactive text.

[0176] The sub-recorded voice refers to the recorded voice corresponding to the applicant, the voice segmentation refers to the small segments of voice obtained by segmenting the sub-recorded voice, and the interactive text refers to the text corresponding to the voice segmentation.

[0177] S332, determining the middle moment corresponding to the speech segment, obtaining the recording frame corresponding to the middle moment as a retrospective frame, and adding the interactive text at a preset position of the retrospective frame to obtain a sub-overview image.

[0178] Specifically, for each speech segment, there is a corresponding recording time period. In the corresponding recording time period, the recorded video corresponding to the speech segment may have multiple recording frames. However, when displaying, in order to quickly preview the question and answer data, one frame can be determined from multiple recording frames, and the interactive text corresponding to the question and answer data can be combined with the recording frame, so that the interviewee can view the image data of the question and answer situation in the corresponding time period, so that the middle moment in the recording time period can be obtained. According to the middle moment, the corresponding recording frame can be found from the recorded video, and the corresponding recording frame is determined as a retrospective frame. The interactive text corresponding to the speech segment is added at the preset position of the retrospective frame. For example, the corresponding interactive text can be added below the retrospective frame to generate the sub-overview image corresponding to each speech segment, and the sub-overview images can be spliced ​​later. Among them, the retrospective frame refers to the recording frame corresponding to the middle moment of the speech segment, the preset position refers to the pre-set position area for displaying the interactive text, and the sub-overview image refers to the overview image corresponding to each speech segment.

[0179] S333: According to the time sequence and the preset arrangement sequence of the speech segments, each of the sub-overview images is spliced ​​to obtain an overview image corresponding to the corresponding backtracking node.

[0180] Specifically, the voice data of the applicant is obtained in chronological order. Therefore, when displaying the question and answer data, the corresponding sub-overview images can be displayed in chronological order. When splicing the sub-overview images, the arrangement order can be set in advance. For example, the sub-overview images can be arranged in order from top to bottom, so that all the sub-overview images can be spliced ​​together in order to form an overview image corresponding to a backtracking node. The question and answer data of the applicant during the interview process can be intuitively viewed through the overview image.

[0181] S34, calling the backtracking template, updating the video display area in the backtracking template based on the interactive segment, updating the overview area in the backtracking template based on the overview image to obtain backtracking data, binding the corresponding backtracking data and backtracking nodes to obtain the interactive backtracking chain.

[0182] See also Figure 3 , which is a schematic diagram of an interactive backtracking chain provided by an embodiment of the present invention, such as Figure 3 As shown in , a backtracking node can be constructed in the middle of the interaction segment. This solution can bind the backtracking template to the backtracking node. By clicking the backtracking node, the backtracking template can be called out. The backtracking template is a template that can display the interaction segment and the overview image. The backtracking template is divided into two parts: the video display area and the overview area. The video display area can display the video data corresponding to each interaction segment, and the overview area can display the overview image.

[0183] By updating the video data corresponding to the interaction segment to the video display area in the retrospective template, you can view the question and answer data corresponding to the candidate in the interaction segment. By updating the overview image to the overview area in the retrospective template, you can quickly preview the candidate's question and answer data, and then you can get the retrospective data corresponding to the candidate. By binding the retrospective data and the retrospective node, you can get the following: Figure 3 The interaction back-tracing chain shown.

[0184] Through the above implementation, the video data of the interview process can be displayed more intuitively, so that the interviewer can quickly trace back the question and answer data of the applicant, thereby improving the tracing efficiency.

[0185] S4, summarizing the interactive backtracking chains of the candidate terminals, obtaining backtracking summary data and sending it to the evaluation terminal.

[0186] It is understandable that during the interview process, more than one candidate may be interviewed, so the interactive backtracking chains corresponding to multiple candidates may be obtained. In order to enable the interviewer on the evaluation side to quickly backtrack and view the interview data of all candidates, the interactive backtracking chains corresponding to multiple candidates may be aggregated and processed to obtain the overall backtracking summary data, so that the interviewer can quickly backtrack. The backtracking summary data refers to the data obtained by aggregating each interactive backtracking chain.

[0187] Through the above implementation, the interviewer can quickly review the interview data of each applicant to improve the interview efficiency.

[0188] See also Figure 4 , is a structural diagram of a digital human resource management statistics system provided by an embodiment of the present invention. The data processing system of the digital human resource management statistics system includes:

[0189] A receiving module is used to receive a screening request from the evaluation end, obtain the application template of each candidate end and input it into the screening training model for question screening, and obtain a customized list of questions corresponding to each candidate end;

[0190] A generating module, configured to generate automatic questioning data corresponding to each candidate terminal according to the customized question list, and obtain recorded video data generated by each candidate terminal based on the automatic questioning data;

[0191] A construction module, used to construct a backtracking node corresponding to each interaction segment in the recorded video data, and generate an interaction backtracking chain corresponding to the candidate end according to each backtracking node;

[0192] The sending module is used to summarize the interactive backtracking chains of each candidate terminal, obtain the backtracking summary data and send it to the evaluation terminal.

[0193] Figure 4 The apparatus of the embodiment shown can be used to perform Figure 1 The implementation principles and technical effects of the steps in the method embodiment shown are similar and will not be repeated here.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A human resource management statistical method based on digitalization, characterized in that: include: Receive a screening request from the evaluation end, obtain the application template of each candidate end and input it into the screening training model for question screening, and obtain a customized list of questions corresponding to each candidate end, specifically including obtaining the information type corresponding to each information column in the application template according to the screening training model, and the estimated duration configured by the evaluation end for each information type, and the information type of the information column in each application template is consistent; Extracting the filled information in each of the information columns in the application template, traversing the question database based on the filled information and the estimated duration, and determining a plurality of selected questions in the question database corresponding to the corresponding information type; Arrange the selected questions from large to small based on the screening weight of each selected question to obtain a list of sub-questions corresponding to the information type, and arrange the sub-question lists according to a preset arrangement order of each information type to obtain a customized list of questions corresponding to each candidate terminal; Among them, traversing the question database based on the filled-in information and the estimated duration, determining a plurality of selected questions corresponding to the corresponding information type in the question database includes: obtaining the screening level selected by the evaluation end, determining the question database corresponding to the screening level, obtaining a sub-database in the question database with the same information type as the filled-in information, the question database including a sub-database corresponding to each of the information types; obtaining the screening time period selected by the evaluation end, determining the storage time of each historical question in the sub-database, obtaining the historical questions with the storage time within the screening time period as a screening question; determining a plurality of information tags corresponding to the filled-in information according to the keyword entries corresponding to the filled-in information, and sorting each of the screening questions based on the information tags. The plurality of customized tags corresponding to the questions are traversed, and the customized tags identical to the information tags are determined to be identical tags; based on the ratio of the identical number of identical tags in each of the first screening questions to the total number of tags, the information matching degree between the filled-in information and each of the first screening questions is obtained, and the first screening questions whose information matching degree is greater than the information matching threshold are obtained as the second screening questions; according to the usage frequency and the response time limit corresponding to each of the second screening questions, the screening weight of each of the second screening questions is determined, and the second screening questions are screened based on the screening weight and the estimated time to obtain the selected questions corresponding to the information type; according to the customized list of questions, the automatic questioning data corresponding to each of the candidate terminals is generated, and the recorded video data generated by each of the candidate terminals based on the automatic questioning data is obtained; Constructing a backtracking node corresponding to each interaction segment in the recorded video data, and generating an interaction backtracking chain corresponding to the candidate end according to each backtracking node; The interactive backtracking chains of the candidate terminals are summarized and processed to obtain the backtracking summary data which is sent to the evaluation terminal.

2. The method according to claim 1, characterized in that Determining the screening weight of each of the secondary screening questions according to the usage frequency and the response time limit corresponding to each of the secondary screening questions, and screening the secondary screening questions based on the screening weight and the estimated time to obtain the selected questions corresponding to the information type, including: Obtaining the usage frequency of each of the secondary screening questions within the screening time period, and obtaining a type of adjustment coefficient according to the ratio of the usage frequency to the benchmark frequency; Retrieving a time limit coefficient list corresponding to the information type, traversing the time limit coefficient list according to the preset response time limit corresponding to the secondary screening question, and determining that the preset coefficient corresponding to the time limit interval where the response time limit is located in the time limit coefficient list is the second type of adjustment coefficient; A comprehensive adjustment coefficient is obtained according to the sum of the first adjustment coefficient and the second adjustment coefficient, and a screening weight corresponding to the secondary screening question is obtained by shifting the reference weight based on the comprehensive adjustment coefficient; The secondary screening questions are arranged from large to small based on the screening weight to obtain a screening sequence, and the selected questions are determined for the screening sequence according to the estimated duration and the response time limit corresponding to each secondary screening question.

3. The method according to claim 2, characterized in that Determining the selected questions for the screening sequence according to the estimated duration and the response time limits corresponding to the secondary screening questions includes: Determine the first secondary screening question in the screening sequence as a candidate question; when the response time limit of the candidate question is less than or equal to the estimated time, determine the candidate question as a selected question; determine the secondary screening question adjacent to the selected question in the screening sequence as the next candidate question; and obtain the remaining time according to the time difference between the estimated time and the response time limit of the selected question; When the reply time limit of the question to be selected is greater than the estimated time, a variable time is obtained according to the time difference between the reply time limit of the question to be selected and the estimated time. When the variable time is less than the time threshold, the question to be selected is determined to be the selected question, and the determination of the selected question is stopped. When the change duration is greater than or equal to the duration threshold, the candidate question is deleted, and the first secondary screening question in the screening sequence is continuously determined as the next candidate question; Repeat the above steps of determining the selected question, and use the remaining time obtained last time as the estimated time when determining the next selected question, and stop determining the selected question when the remaining time is less than or equal to the time threshold.

4. The method according to claim 1, characterized in that: Generating automatic questioning data corresponding to each candidate terminal according to the customized question list, and obtaining recorded video data generated by each candidate terminal based on the automatic questioning data, including: Obtaining question texts corresponding to each selected question in the question customization list, generating question voices according to the question texts, and generating automatic question videos corresponding to the selected questions based on the virtual digital human and the question voices; Arrange the automatic questioning videos according to the arrangement order corresponding to the selected questions to obtain an automatic questioning sequence corresponding to the candidate terminal, wherein the automatic questioning data includes the automatic questioning sequence; Switching to active communication mode based on the interview connection information of the candidate terminal, automatically displaying the automatic questioning video according to the automatic interaction area in the interaction interface, and displaying the active answering video of the candidate terminal based on the real-time interaction area in the interaction interface; Switching to the passive communication mode according to the end information of the active communication mode, displaying the passive question video of the candidate terminal based on the real-time interaction area, and displaying the passive answer video corresponding to the question video based on the automatic interaction area; The video data in the interactive interface in the active communication mode and the passive communication mode are recorded to obtain the recorded video data.

5. The method according to claim 4, characterized in that Switching to the active communication mode based on the interview connection information of the candidate terminal, automatically displaying the automatic questioning video according to the automatic interaction area in the interaction interface, and displaying the active answering video of the candidate terminal based on the real-time interaction area in the interaction interface, including: Selecting the automatic questioning videos in the automatic questioning sequence as target questioning videos in sequence, displaying the target questioning videos based on the automatic interaction area, and obtaining the active answering videos of the candidate terminals in the real-time interaction area after the display is completed; When the response time of the active answer video is equal to the response time limit of the target question video, or when a switching keyword is extracted based on the answer voice of the active answer video, the next target question video is displayed; Switching to the passive communication mode according to the end information of the active communication mode, displaying the passive question video of the candidate terminal based on the real-time interaction area, and displaying the passive answer video corresponding to the question video based on the automatic interaction area, including: When there is no next target question video, responding to the end information of the active communication mode and switching to the passive communication mode; The passive question video of the candidate terminal in the real-time interaction area is obtained, and the corresponding passive answer video is retrieved according to the question keyword corresponding to the passive question video and displayed in the automatic interaction area.

6. The method according to claim 5, characterized in that Constructing a backtracking node corresponding to each interaction segment in the recorded video data, and generating an interaction backtracking chain corresponding to the candidate end according to each backtracking node, including: Acquire the video segments of the automatic questioning video and the active answering video corresponding to the recorded video data, and the video segments of the passive questioning video and the passive answering video corresponding to the recorded video data as interaction segments; Constructing a recording timeline corresponding to the recorded video data, determining a recording segment corresponding to each of the interaction segments in the recording timeline, and constructing a backtracking node corresponding to each of the interaction segments at a center position of each of the recording segments; Generate an overview image corresponding to each backtracking node according to the recorded frames and recorded voices corresponding to each of the interaction segments; Retrieve a backtracking template, update a video display area in the backtracking template based on the interactive segment, update an overview area in the backtracking template based on the overview image to obtain backtracking data, bind the corresponding backtracking data and backtracking nodes to obtain the interactive backtracking chain.

7. The method according to claim 6, characterized in that Generating an overview image corresponding to each backtracking node according to the recorded frames and recorded voices corresponding to each of the interaction segments, including: Obtaining a sub-recorded voice corresponding to the candidate end in the recorded voice, performing voice segmentation recognition on the sub-recorded voice to obtain voice segments, and extracting interactive texts corresponding to each of the voice segments; Determine an intermediate moment corresponding to the speech segment, obtain a recording frame corresponding to the intermediate moment as a backtracking frame, and add the interactive text at a preset position of the backtracking frame to obtain a sub-overview image; According to the time sequence and the preset arrangement sequence of each of the speech segments, each of the sub-overview images is spliced ​​to obtain an overview image corresponding to the corresponding backtracking node.

8. A statistical system for the digital human resource management statistical method according to claim 1, characterized in that: include: A receiving module is used to receive a screening request from the evaluation end, obtain the application template of each candidate end and input it into the screening training model for question screening, and obtain a customized list of questions corresponding to each candidate end; A generating module, configured to generate automatic questioning data corresponding to each candidate terminal according to the customized question list, and obtain recorded video data generated by each candidate terminal based on the automatic questioning data; A construction module, used to construct a backtracking node corresponding to each interaction segment in the recorded video data, and generate an interaction backtracking chain corresponding to the candidate end according to each backtracking node; The sending module is used to summarize the interactive backtracking chains of each candidate terminal, obtain the backtracking summary data and send it to the evaluation terminal.

Images