KR20200145299A - Intelligent recruitment support platform based on online interview video analysis and social media information analysis - Google Patents

Abstract

The present invention relates to an online interview system providing an interview video in which idle time unnecessary on an interviewer′s part is removed. The system includes: an interview terminal on which an interview applicant receives question information from a service server, forms video data in which an interview video corresponding thereto is filmed, and transmits the data; the service server receiving and storing the video data through a network, processing and editing the video data, and transmitting the processed and edited video data to an interviewer terminal; and the interviewer terminal receiving the processed and edited video data from the service server through the network. The system includes: a signal input unit receiving the input of a talent information request signal related to a talent to be discovered; a talent information collection unit tracking open information on the talent′s past and present and receiving talent-related big data corresponding to a keyword and including data having continuity and relevance to related work; a talent information processing unit converting the talent-related big data from online data and analyzing the talent-related big data to generate talent status information corresponding to the talent information request signal; a talent information storage unit storing the converted talent-related big data; and a talent information output unit outputting the talent status information.

Description

Intelligent recruitment support platform based on online interview video analysis and social media information analysis}

The present invention relates to an online interview video analysis and social media information analysis-based intelligent recruitment support platform that analyzes video through an online interview and generates data for evaluating talent based on social media information analysis. .

Recently, companies' hiring methods for new employees have changed a lot from documents and written examinations to interview evaluation and aptitude evaluation. This is because the interviewer and the interviewer can comprehensively grasp or express personality and accidents through the interview. In addition, in a situation where the need for creative and active talents is emphasized according to various business environment changes, the importance of interviews as a way to grasp them is growing more than before.

In the conventional case, a number of applicants are required to register their résumés, information on the type of business that they wish to find a job for, and region information on the website, and are provided with employment information from a number of recruiting companies and provide job information linked to each interviewer. .

Applicants who have passed the document interview from the website will visit the job offer company directly and receive an offline interview to decide whether to hire. However, the applicant must visit the recruitment company in person, and if the distance between the applicant and the recruitment company is far away, the inconvenience caused by the movement becomes considerable. In addition, since the interviewees have to prepare for offline interviews for applicants regardless of the number of recruitment scales or small, there is a considerable burden in terms of time and cost.

Therefore, it is necessary for the interview applicant and the interviewer to save time and cost, and to have an excellent interview effect. With the development of communication technology, video interviews by video and audio between remote terminals have emerged as an alternative. This is a form that basically consists of an interviewer's terminal and an interviewer terminal that registers a video through it and receives and checks the registered video, and a number of related patent applications have been made.

In Korean Patent Publication No. KR10-2004-0006879, it provides job information for each job seeker to a number of job seekers, and makes it easy to produce interview videos or self-introduction videos for job search and job search, and provide them in real time to a number of job seekers. The video interview applause and the online interview system using it are being launched.

Registered Patent No. KR10-0453838 discloses the configuration of an overall online interview system for online job search, job search processing steps, and online interview performance and information management.

In Korean Patent Application Publication No. KR10-2002-0069973, information of a number of companies or organizations seeking job offers from a server connected through an Internet network is collected, and applications, resumes, self-introductions, and interviews are provided to a number of individual users seeking employment. It discloses a video resume and video interview service method using the Internet to produce and transmit the contents of. In addition, patents that have a configuration that adopts various methods of evaluating interview videos are also being applied.

Meanwhile, most of the patents related to video interviews are provided in the form of receiving pre-entered questions as a video or text screen, and responding to them. In this case, the video of the interview applicant includes the content of the interview question and the time required for grasping the question, such as handling time.

There is also a section in which the interview applicant is not speaking during the time it takes to answer the question. That is, so-called idle time (idle time) occurs. For that reason, unnecessary time can be wasted in the position of the interviewer who receives and reviews the video of the interviewer.

In the specific process of the video interview, if a certain answer time is provided to the evaluation questions for the interview applicant, if the answer time cannot be completed, idle time may occur in the process of answering.

Idle time may occur even while inputting a signal indicating that the answer has been completed. When a question is provided by categorizing evaluation content by evaluation item, unnecessary idle time may occur from the perspective of an actual interviewer even in the process of explaining the evaluation content.

In recent years, due to the activation of the authentication means, a request for a question to the actual applicant interviewee is processed through SMS of a mobile phone, or an identity verification procedure is performed by pressing the authentication number of the SMS. These times can be idle time in the process of evaluating the interviewer's interview video from the position of the interviewer.

From the standpoint of a company or institution that is seeking a job, the contents of the answers that are far from the question or the contents of the answers that do not include important keywords may be unnecessary, and the video section including them may be unnecessary review time from the viewpoint of the interviewer.

Conversely, the video section including a specific keyword may be an important video section from the perspective of the interviewer. On the other hand, from the standpoint of companies or institutions that are seeking employment, it may be necessary to find and exclude applicants for interviews that are below the level more quickly, and for those who require additional in-depth interviews or offline interviews, video interview videos are more closely related. You may also need to consider

In the end, for the efficiency of the video interview, it is necessary to provide an interview video with unnecessary idle time removed from the position of the interviewer. Alternatively, for efficiency and accuracy of the interview, specific sections need to be searched again or interview data in a more summarized form needs to be provided.

The present invention provides an interview system that provides an interview video with unnecessary idle time removed from the position of the interviewer, and analyzes an online interview video that can extract and provide video parts necessary for accuracy and efficiency of the interview evaluation. And an intelligent recruitment support platform based on social media information analysis.

The intelligent recruitment support platform based on online interview video analysis and social media information analysis according to the present invention includes: an interview terminal for receiving question information from a service server, forming video data corresponding to the interview applicant, and transmitting the video data; A service server receiving and storing the moving picture data through a network, processing and editing the moving picture data, and transmitting the processed edited moving picture data to an interviewer terminal; And an interviewer terminal receiving the processed and edited video data from the service server through a network, wherein the keyword related to the person to be discovered and the research field, work field, name, region, and age of the person to be discovered , A signal input unit for receiving a signal for requesting human resources information including one or more of career, achievement, patent, work place, and school from which they graduate; A site that includes at least one of thesis search site, social network service post, blog post, personal dictionary site, newspaper article, social networking site, university site that provides university alumni information, specialized report site for each field, and public patent search site A human resource information collection unit that tracks the past and present information of the human resource from the human resource and receives big data related to the human resource corresponding to the keyword including the relevant data and continuity of the related work; Related to the talent in a form that can be searched through one or more of the keywords included in the talent information request signal, the research field of the talent, work field, name, region, age, career, achievement, patent, work place, and school A talent information processing unit that converts big data and analyzes the talent-related big data to generate talent status information corresponding to the talent information request signal; A talent information storage unit that stores the converted big data related to the talent; And a talent information output unit that outputs the talent status information, wherein the talent information output unit includes the keyword included in the talent information request signal, the research field of the talent, the field of work, the name, region, age, career, achievement, It visualizes the human resource status information in the form of a Venn diagram according to one or more criteria among patents, workplaces, and graduated schools to subdivide and output the distribution of human resources.Keywords related to the human resources to be discovered, research fields, work fields, and Receiving a signal for requesting talent information including one or more of name, region, age, career, achievement, patent, work place, and school from the signal input unit; A site that includes at least one of thesis search site, social network service post, blog post, personal dictionary site, newspaper article, social networking site, university site providing university alumni information, specialized report site for each field, and open patent search site Searching for, tracking past and present information of the talent, and collecting, at a talent information collection unit, talent-related big data corresponding to keywords containing relevant data and continuity of related work; It can be searched through one or more of the keyword, the research field of the talent, the field of work, the name, the region, the age, the career, the achievement, the patent, the place of work, and the school from which it is a graduate, included in the request signal for the talent information received from the signal input unit Converting the human resource-related big data into an existing form and storing the converted human resource-related big data in a talent information storage unit in a talent information processing unit; And generating talent status information corresponding to the talent information request signal by analyzing the talent-related big data stored in the talent information storage unit by the talent information processing unit, and outputting the generated talent status information from the talent information output unit. Including, and the step of outputting the information comprises the keyword included in the talent information request signal, the research field of the talent, work field, name, region, age, career, achievement, patent, work place, one or more of the school from which The human resource status information visualized in the form of a Venn diagram that subdivides the distribution of human resources by standard is output, and the service server detects voice data of video data and converts the detected voice data into text data. A data matching unit that indexes and indexes the video data corresponding to the indexed text data to match text data and video data; A data determination unit for evaluating the characteristics of the text data; A data extracting unit for selecting the text data according to the evaluation of the data discriminating unit and extracting moving image data corresponding to the index of the selected text data; And a data merging unit for merging the moving image data obtained through the data extracting unit, wherein the interviewer terminal includes a data editing request unit for requesting processing and editing of the moving image data so that the moving image section including a specific keyword is included in the service server. Including, wherein the method of indexing the text data in the data matching unit indexes time as a function, but is performed in word units of the text, and the data discrimination unit evaluates the interval of time indexed in units of words as a function to determine idle time. Is evaluated, and also evaluates whether a specific keyword is included in the text data, and the data extraction unit selects text data for extracting video data, excluding the part evaluated as the idle time part of the text data, Text data including keywords is selected in units of sentences including words including specific keywords, and moving image data corresponding to the index of the selected text data is extracted.

The present invention provides an interview system that provides an interview video with unnecessary idle time removed from the position of the interviewer, and analyzes an online interview video that can extract and provide video parts necessary for accuracy and efficiency of the interview evaluation. And an intelligent recruitment support platform based on social media information analysis.

1 is a diagram showing a human resource information processing apparatus maintaining connection with various sources capable of collecting human resource information according to an embodiment of the present invention;
Fig. 2 is a block diagram illustrating the human resource information processing device of Fig. 1;
FIG. 3 is a flow chart illustrating a process of generating talent-related information by collecting and processing information by the talent information processing apparatus of FIG. 2;
4 is a diagram showing a process of extracting and analyzing talent-related big data;
5 is a diagram showing an example of human resource status information generated and output according to an embodiment of the present invention;
6 is a diagram showing another example of human resource status information generated and output according to an embodiment of the present invention;
7 is a basic configuration diagram for explaining the online interview system of the present invention;
8 is a configuration diagram of an interview terminal of the present invention,
9 is a configuration diagram of a service server of the present invention;
10 is a block diagram of the interviewer terminal of the present invention.

The present invention can be embodied as computer-readable codes on a computer-readable recording medium, and the computer-readable recording medium includes all types of recording devices storing data that can be read by a computer system. . Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage devices. Further, the computer-readable recording medium is distributed over a computer system connected by a network, so that the computer-readable code can be stored and executed in a distributed manner.

All terms used have the same meaning as commonly understood by one of ordinary skill in the field to which the present invention belongs, unless otherwise defined. Terms defined in commonly used dictionaries should be construed as having meanings in the context of related technologies, and cannot be construed as having an ideal or excessive formal meaning unless explicitly defined in the present application.

Before describing the embodiments of the present invention, terms and concepts to be used hereinafter will be briefly described. The semantic web is a machine, that is, a computer that can process information about resources (web documents, various files, services, etc.) and the relationship-semantic information (semanteme) between resources in a distributed environment like the current Internet. It is a framework and technology that expresses it in the form of ontology and allows automated machines to process it.

Ontology is a model that abstracts and shares what people think about things, and refers to a technology in which the types of concepts and constraints on use are explicitly defined and standardized. In other words, ontology is a technology that enables logical reasoning, such as a computer understanding and processing the concept of words.

In the field of computer science, an ontology is a data model representing a specific domain and is defined as a set of formalized vocabularies that describe concepts belonging to a specific domain and the relationship between concepts. In particular, Ontology is a tool that can implement a semantic web and is used as a tool that can semantically connect knowledge concepts.

Web information search made it possible to search for all electronic resources that can be accessed through the web. Ontology, in particular, makes it possible to process the concept of a person's object in a computer by processing it into a kind of database. Due to the rapid development of the Internet, as the scope of search targets expanded, more sophisticated search was required, and it was an opportunity to promote the development of an intelligent information search system. Based on this opportunity, the need for a new tool for information retrieval that can effectively manage web resources has emerged, and Ontology has been in the spotlight.

Ontology is used in the fields of machine translation of natural language and artificial intelligence, and recently, it has been attracting attention as a core element such as the semantic web using ontology that describes the relationship between Internet resources in a specific field and the semantic web service derived from it. have.

If the environment that the machine infers information according to the meaning and provides it to the user is called the Semantic Web, among them, the framework that helps the machine interpret the meaning can be called an ontology. Ontology in information technology refers to a working model of the interaction between'substance' and'what the substance means' (or data and the'meaning' that the data means) within a certain domain. It needs a lubricant between the and virtual (its web representation). It can be considered as a collection of standards and best practices for sharing on the web.

Semantic Web technology expresses understanding of language in human head in computer language and makes it usable by computer. It is implemented on the web with special distributed environment. This is to make the machine better understand and respond appropriately to human needs, such as IR.

For true communication between man and machine, machines must be able to understand language at the level that humans can understand. Conceptualization is a method of grasping and understanding the features of each object or event in the world while experiencing each object or event in the world. Ontology can be said to be a technology that makes the same concept as a person in a computer into a kind of database. Ontology, designed to help programs and humans share knowledge, aims to find more accurate information by clearly defining the concept of the resource that is the target of the information system and describing it in detail.

The semantic web can be expressed as a semantic markup language based on XML (Extensible Markup Language). In this semantic web, the concept is expressed in the form of a triple of a subject, a predicate, and an object, and again, each subject, predicate, and object can be expressed as a URI (Uniform Resource Identifier) of XML. I can. Currently, standard languages for describing semantic web ontology include RDF, RDF-Schema, OWL, SW RL and TopicMaps proposed by ISO.

Components of an ontology can be classified into class, instance, relationship, and property. It can be explained that a class is generally a name that we attach to things or concepts.

Instance refers to the concrete object or event itself of an object or concept that appears in a practical form. Class and instance classification can be very different depending on the application, purpose of use, and situation. In other words, an object with the same expression can become a class in some cases and an instance in other cases. In addition, a property is a connection of a class or instance with a specific value to indicate a specific property or propensity of a class or instance.

Relation refers to relations that exist between classes and instances, and can generally be classified into taxonomic relation and non-taxonomic relation. The taxonomic relation is a relation that is expressed hierarchically by dividing into higher concepts and lower concepts for conceptual classification of classes and instances. And the non-taxonomic relation refers to a relationship that is not a classification relationship (Taxonomic Relation). For example, "being healthy from exercise" is expressed using a cause-result.

1 is a diagram illustrating a human resource information processing apparatus maintaining connection with various sources capable of collecting human resource information according to an embodiment of the present invention.

Referring to FIG. 1, the human resource information processing apparatus 100 may collect human resources-related information from various sites. Here, the human resource information processing device receives and collects information related to human resources in connection with general portal sites, social network service sites, domestic and international thesis search sites, job search sites, and various sites that provide professional reports for a fee or for free. I heard. Here, social network services may include Facebook, Instagram, LinkedIn, and the like, and there is no limitation on the type.

There is no restriction on a site to which the human resource information processing apparatus 100 can be connected or a server thereof, and there is no particular restriction on the content or type of information that can be received from the site or server. If the paper search site or report search site to be linked is a paid site, a step of paying for it in advance may be added. In addition, the human resource information processing device 100 maintains a connection with one or more of various blogs, homepages, Internet personal dictionary, Internet newspaper articles, social networking sites, university sites that provide university alumni information, and open patent search sites. While receiving information related to talent.

2 is a block diagram illustrating the human resource information processing apparatus 100 in FIG. 1.

Referring to FIG. 2, the human resource information processing apparatus 100 includes a signal input unit 110, a human resource information collection unit 120, a human resource information processing unit 130, a human resource information storage unit 140, and a human resource information output unit 150. It may include.

The signal input unit 110 may receive a human resources information request signal for requesting information on human resources to be discovered. Such a talent information request signal may be input in the form of receiving a keyword for searching for talent. The keyword input itself may be recognized as a talent information request signal, or a separate talent information request signal may be input after keyword input and search.

Search for author information such as past research results (papers, journals) in special research fields within science and engineering or promising technologies, and search for the work field and related social network services that the authors are currently engaged in, and if there is continuity of related work Manpower may be registered in the talent pool according to an embodiment of the present invention, and the corresponding manpower information may be matched with the personnel information of a job-related company to be used as basic data such as headhunting or talent invitation.

Although there are known technologies that have been partially mentioned for a human resource database based on social network services or other information, and a job search and job search program based on this, the relevant manpower is not subject to past professional research projects or A technology that tracks all basic information such as papers, journals, etc. that occurred when performing work, as well as contents posted through the currently open social network service and related tasks, and compares them with relevant data, has not yet been disclosed.

As mentioned above, the talent information request signal may include a keyword for a talent to be discovered or hired. In addition to the keyword, the information may include one or more of the research field, work field, technology field, name, region, age, career, achievement, patent, work place, and school from which the recruitment screening target is performed. Here, the region information may be information on the place of origin, work place, research place, and residence of the relevant person. More detailed information, such as whether the person is a master's or doctoral level, how many papers have been published in major journals, what is the paper, has a patent, and if so, what the patent number is, can be included in the Talent Information Request signal. have.

Even if such information is not included in the human resource information request signal, according to an embodiment of the present invention, it may be possible to collect and manage such information together in a data collection and analysis process to be described later.

As mentioned above, the talent information collection unit 120 may extract and collect talent-related information from various sites or servers thereof. In particular, in the case that the relevant person has published a thesis or applied for a patent, the human resource information collection unit may store the entire thesis or patent document as human resource-related data, and these data can store the entire document. In other words, talent-related information can be a fairly large amount of data.

For example, in the case of collecting talent-related information from posts based on social network services, all posts related to the job or major of the person in question may be searched and received through the talent information receiving unit 120. In this case, the information related to the talent is big data Can have the form of. In the case of the manuscript, as well as the case where the relevant manuscript is the main author of the manuscript, as well as the case where it is a sub-author, if it is used as a reference for another manuscript, all the corresponding data can be transmitted to the talent information processing device 100 as talent-related information. .

The talent information processing unit 130 is in a form that can be searched through one or more of a keyword for discovering or recruiting talent, research field of talent, work field, name, region, age, career, achievement, patent, place of work, and school of origin. It is possible to transform the above talent-related big data. In addition, the talent information processing unit 130 may analyze the talent-related big data to generate talent status information corresponding to the talent information request signal.

When the talent information processing unit 130 converts the collected talent-related big data, Hadoop, HBASE, NoSql (Cassandra), or the like may be used. Hadoop is a technology that processes large amounts of data by grouping several inexpensive computers as if they were one. Hadoop is a distributed file system that provides the ability to store large files on thousands of distributed x86 devices, and MapReduce, a computing platform that allows you to quickly and easily analyze stored file data using the CPU and memory resources of the distributed server. It is composed. With Hadoop, large amounts of data can be processed quickly and at a relatively low cost.

In addition, the talent information processing unit 130 may collect talent-related data from a data source using a plum for data collection and store it stably in the Hadoop file system. In addition, the human resource information processing unit 130 may perform a search for big data related to human resources using elastic search. In addition, the human resource information processing unit 130 structurally stores large-capacity data related to human resources in a distributed server using HBASE or Cassandra, thereby enabling high-speed real-time storage and providing a structured data storage equipped with an inquiry function.

The fact that the talent information processing unit 130 analyzes talent-related information or talent-related big data means that by employing an ontology-based big data analysis method, it is possible to quickly analyze whether the job-seeker-related data meets the conditions desired by the job seeker. For example, when a recruiter enters a keyword to obtain talent information, the more the input keyword is added, the talent information processing device 100 can accurately grasp the talent information desired by the job seeker, and noise is removed, resulting in higher accuracy and higher user accuracy. It will be possible to create and provide information on the status of talent that meets the needs of the company.

The talent information processing unit 130 generates talent status information corresponding to a request from the talent information request signal, based on the analyzed big data related to talent.

Talent status information may be information indicating how a talent pool in a specific field is formed. Depending on the degree or experience, how many master's and doctoral level personnel in a specific field are, how many professional personnel in their 30s and 40s are each, in which region, gender, and who are each of those personnel, etc. This talent status information can be included. In other words, the human resource status information includes at least one of the keywords, research field, work field, name, region, age, career, achievement, patent, work place, school of study, and job search information included in the previously inputted talent information request signal. It includes, and may be information classified or sorted by one or more of these criteria.

Talent status information is the human resource that is derived from the search results when searching by keyword, specialized research field, work field or detailed major, the age group or major career of the manpower, where he is currently working, and what job he is currently looking for. May include information about.

Talent status information may be output as information in the form of a graph visualized as a band diagram, a bar graph, or a band graph. This will be additionally described later in FIGS. 5 and 6.

The human resource information storage unit 140 of the human resource information processing apparatus 100 stores the converted human resource-related information or talent-related big data. In addition, the big data related to talent stored in the talent information storage unit 150 may be periodically updated.

The human resource information output unit 150 outputs human resource status information. As described above, the talent status information may provide readable information on the distribution of talent or the current situation of each talent, and for this purpose, it may be information visualized in the form of a graph.

FIG. 3 is a flowchart illustrating a process of collecting, processing, and outputting human resource information by the apparatus for processing human resources according to an embodiment of the present invention described in FIGS. 1 and 2.

The signal input unit 110 of the human resource information processing apparatus 100 receives a human resource information request signal (S210). As described above, the signal for requesting talent information can be a keyword to search for the talent to be discovered, and the research field, work field, detailed major, name, region, age, career, achievement, patent, work place, and origin of the relevant talent. It can include information about school, job search, etc.

When the signal input unit 110 receives a signal for requesting talent information, the talent information collection unit 120 provides a general portal site, a social network service site such as Facebook or Instagram, LinkedIn, a domestic and international paper search site, and a job search specialist. Human resources-related information can be collected by receiving all information and data related to the relevant human resources from various sites that provide sites, professional reports for a fee or for free, or from the server (S220). Here, the information related to human resources may be a large amount of big data.

Subsequently, the collected information may be converted and stored in a form that can be retrieved later (S230), and in particular, various tools necessary for transformation and analysis of big data may be used. Here, for example, Hadoop, Cassandra, HBASE, etc. can be used.

When the human resource information processing unit 130 analyzes the converted human resource-related information to generate the human resource status information, a big data analysis method based on ontology may be used. According to the big data analysis method based on ontology, the keyword entered by the user is not confused or mixed with terms in other fields, and by inferring the user's needs by the combination of keywords, the user's intended job seeker can be extracted more accurately. You will be able to. Through this process, the talent information processing unit 130 may generate information on the status of talent (S240).

The talent information output unit outputs the talent status information generated above (S250). The human resource status information output here may be a list of personnel to be screened for hiring or not, or may be in the form of a graph to identify the distribution of personnel, career details, job search status, and current work.

4 is a diagram illustrating a process of collecting and processing talent-related big data by an ontology-based big data analysis method according to an embodiment of the present invention.

In the case where the person in charge of recruitment intends to employ a professional manpower related to “3D video” through the apparatus or method for processing human resource information according to an embodiment of the present invention, a process of deriving a result through an ontology-based analysis method is illustrated as an example.

When a user inputs only “3D” as a search word, it is unclear whether it is a search request for 3D printing, a search request for 3D image, or a search request for 3D hologram. This means that when searching for “3D video” again, search results irrelevant to the user's original intention can be automatically removed.

In the 3D image field, detailed fields are searched together, and “3D image conversion”, “3D image generation”, and “3D image display device” are searched together, and experts in each detailed field are derived as a result. Accordingly, according to an exemplary embodiment of the present invention, the human resource information processing apparatus 100 may provide a search result for a manpower that more accurately meets a user's request, as well as a result for a more detailed search. In addition, the human resource information processing device 100 generates accurate and rich human resource status information by storing and analyzing a vast amount of data in a short time, such as all thesis, articles, reports, social network service posts, etc., searched for each human resource. So that it can be provided.

5 and 6 are diagrams illustrating a state in which human resource status information is output according to an embodiment of the present invention.

FIG. 5 shows the distribution of domestic and foreign experts in a specific business field, and FIG. 6 shows the distribution of domestic experts. For example, assuming that Figs. 5 and 6 are human resource status information indicating the distribution of the working status of domestic and foreign experts on cancer cell tracking technology, the first Venn diagram in Fig. 5 is the workforce of general companies 310 and research institutes in the European region. It shows the distribution of (320) and university workforce (330).

The second Venn diagram of FIG. 5 may represent the distribution of general corporate workforce 330, laboratory workforce 340, and university workforce 350 in Asia with respect to workforces in the same field. Similarly, the third Venn diagram of FIG. 5 may represent the distribution of general corporate workforce 370, university workforce 380, and research institute workforce 390 in North America with respect to manpower in the same field.

By varying the location and size of each Venn diagram, the number of professional personnel distributed in each region can be visually expressed.In each Venn diagram, the distribution of personnel can be subdivided according to criteria such as age, career, gender, and workplace. Can be indicated.

Talent status information indicating the distribution of domestic specialists shown in FIG. 6 may also be interpreted as well.

As illustrated in FIG. 6, when comparing the Seoul Gyeonggi area, Daejeon, and Gwangju, the number of distributions of professional personnel can be compared at a glance by comparing the size of each Venn diagram. In addition, each Venn diagram can be subdivided into information such as work place, age, gender, career, major, and degree.

Assuming that the information on the current status of domestic talent shown in FIG. 6 is also subdivided according to the type of work, the professional manpower living in Korea in the Seoul-Gyeonggi area for cancer cell tracking technology is the university working manpower 410, through the first Venn diagram of FIG. It can represent the distribution of the labor force at the research institute 420 and the workforce at the general company 430.

For the cancer cell tracking technology, the domestic residents in the Daejeon and Chungcheong regions can show the distribution of the university workforce (450), the research center workforce (440), and the general company workforce (460) through the second Venn diagram in FIG. have. Similarly, the distribution of the university working manpower 470 and the research institute working manpower 480 through the third Venn diagram of FIG. 6 may be indicated for the domestic professional manpower in the Gwangju area for the cancer cell tracking technology.

7 is a basic configuration diagram for explaining the online interview system of the present invention.

The online interview system of the present invention includes an interview terminal 100, a service server 200, and an interviewer terminal 300, and is connected through a network centering on the service server 200.

The service server 200 transmits a question for an interview to the interview terminal 100. The interview applicant receives question information or instructions from the service server through the interview terminal 100, and accordingly shoots an interview image to form video data.

This basic configuration is a general configuration and flow of an online interview system, and the present invention also has such a configuration. It is obvious that the service server 200 can request not only questions through the interview terminal 100 but also items generally requested during the job search process, such as essential personal information and history of the interviewer.

In addition, contents such as allowing the interviewer to view these data are generally self-evident facts from the standpoint of a person skilled in the art, and the specific screen composition, question system, input system, and detailed flow chart related to this can be configured in a wide variety. Matters are omitted so as not to obscure the core technical configuration of the present invention.

The service server 200 stores the video data received from the interview terminal 100 and transmits it to the interviewer terminal 300. The interviewer terminal 300 receives video data and performs an interview evaluation. The interviewer terminal 300 transmits a questionnaire for an interview or various items mentioned above to the service server 200 to cause the service server 200 to display various items on the interview terminal 100. This process can also be said to be a general flow chart of the video interview system.

The core technical idea of the present invention is how to process and edit the video data received from the interview terminal 100 and transmit it to the interviewer terminal 300, thereby securing the efficiency and accuracy of the interview.

8 is a block diagram of an interview terminal according to the present invention.

The interview terminal 100 may basically have a control unit 110, a communication module 120, a display module 130, a video module 140, a camera module 150, and a storage module 160, as required. Different module parts can be configured. The control unit 110 coordinates and controls the functions of each module. Each of the modules can adopt known types.

The interview terminal 100 may be a terminal installed in a preset place, a desktop computer with a program having a platform of the interview system installed, or a mobile terminal such as a portable smartphone, and a plurality of the interview terminals are possible.

The overall interview system is preferably a platform, and such a platform will be installed in the interview terminal 100 to be interlocked with the service server. The video data generated by the interview terminal 100 includes video data and audio data, and the video data and audio data may be indexed to correspond to each other.

The index method may be indexed through time, or may be indexed using various coding methods. It is obvious that the video data and the audio data may be transmitted to the service server 200 in a combined form, or may be transmitted separately, and the combined form may be separated within the service server 200.

The video data sent from the interview terminal 100 to the service server 200 is not only the part of the answer to the question, but also the content of the interview question or the handling time required to understand the question. There may be times when they are included or not responding, and these times may include idle time (idle time).

Various types of idle time occur even if only the answer part is photographed by instructing the start and end of the photographing time through the control unit 110 of the interview terminal 100. Even when the service server 200 sets an answer time for an individual question and uses a method of photographing only in this case, various types of idle time may also occur.

Even if the method of the platform is improved to control or control the flow of the interview terminal 100, in the end, the video data sent from the interview terminal 100 to the service server 200 must include various types of idle time. none. It is possible to improve the platform method to have a minimum idle time, but in this case, the interview applicant must perform various actions other than the interview response, which prevents concentration on the interview, rather, through the service server 200 It would be desirable to edit the data.

The form of the platform will be a form in which a certain question is displayed on the screen as an image or text, and an answer is given for a certain period of time and proceeds to the next question. Of course, various types of platform design will be possible, but the fact that idle time still occurs will not change.

9 is a configuration diagram of a service server according to the present invention, which is a core part of processing and editing moving image data. The service server 200 of the online interview system of the present invention includes a data matching unit 210, a data determining unit 220, a data extracting unit 230, and a data merging unit 240.

The data matching unit 210 detects voice data from the received moving picture data, converts the detected voice data into text data, and indexes the text data, the index being linked with the moving picture data corresponding to the text data. Should be indexed. By doing this, text data and moving image data are matched in a relevance. The data matching unit 210 may include more specific functional units 211 and 212 for performing the above process.

For example, in moving picture data including video data and audio data, it may include a function unit for detecting only an audio portion and converting it into text data. Modules for performing these functions can utilize an open source for detecting voice data (for example, FFmpeg library), and various open sources for voice recognition (for example, HTK open source). It is not specifically mentioned.

There may be various methods of indexing the text data. Perhaps the most convenient way is to index time as a function. By indexing a corresponding time for text data, and mapping or mapping the corresponding index to correspond to the index of the moving image data, the text data and the moving image data are matched with each other.

Therefore, it is possible to apply the processing editing form of text data to moving image data. In the case of indexing time as a function, the index may be in units of syllables or words. That is, a syllable at a certain position of text data has an index of a specific time, and the position of the moving image data to be processed and edited can be detected through the index of the moving image data corresponding to the index.

The data discrimination unit 220 is a part that evaluates various characteristics of text data. In the present invention, the standard for processing and editing moving image data begins with grasping the characteristics of text data. For processing and editing video data, it is said that it is possible to analyze the psychological confidence or emotional state by analyzing the voice by acoustic phonetics, but the possibility of error is high and objective standards have not been established.

Or, for processing and editing video data, it is possible to analyze the motion, facial expression, posture, etc. of the interview applicant in the video and analyze it according to specific criteria, but this is also said to be possible with errors and objective criteria. I think it is difficult to build.

In the present invention, it is also a technical problem to reduce the time for the interviewer to conduct an interview evaluation by performing processing editing to exclude the moving picture of the idle time portion through grasping the characteristics of text data. In addition, it is possible to perform processing editing to search for a part containing a specific answer through the identification of the characteristics of text data, or processing and editing only key points of answers to interview questions in the form of highlights. In addition, it is possible to conduct interview evaluation through a method of systematically assigning scores for evaluation items through characterization of text data.

The data determination unit 220 may analyze the text data to evaluate the idle time portion. Depending on how you define idle time within the platform, you can evaluate a wide variety of idle times. The easiest thing to think of may be the time between question and answer. In addition, the time at which the question is started may also be idle time for the interviewer. This is because the interviewer already knows the content of the question, so that part of the video data may be unnecessary.

If the platform is designed to limit the answer time for a specific question, if the answer has already been completed while the answer time has not passed, the time without answering may also become idle time. As mentioned above, when text data is indexed as a function of time, such idle time can be found by evaluating the time interval between each index as a function.

Of course, it would be possible to find the idle time through several other algorithms. For example, in the analysis or evaluation of text data, it may be evaluated by using sentence units or word units as a function. Since sentences are basically understood as units that contain specific thoughts, the interval between sentences can be idle time. Even in the case of a word, it may contain words that contain thoughts, so if the method of word analysis is objectified through more research, it will be a sufficiently usable function.

The data extracting unit 230 may extract a specific part of the moving picture data according to the evaluation of the data discriminating unit 220 and function to select or discard the extracted moving picture data as necessary. As the premise for extracting a specific part of video data is to evaluate or analyze text data, text data will be extracted first in order, and the video data by searching the index of the video data corresponding to the index of the extracted text data. Will extract.

Taking the above-mentioned evaluation of idle time as an example, video data corresponding to a part corresponding to the idle time will be discarded, or video data corresponding to a part corresponding to a non-idle time part will be extracted.

The data merging unit 240 plays a role of constructing new moving picture data by combining the moving picture data obtained through the data extracting unit 230. Of course, it is possible to configure the merge module by using an open source for data merging, and such matters are omitted so as not to obscure the subject matter of the invention.

Other service servers may include various functional units. It may include a DB that codes for the extracted and merged new video data and stores it under a new name, and other elements for various platform functions that can be implemented for the convenience of an online interview system. Yes, of course.

10 is a block diagram of the interviewer terminal 300 of the present invention, and may include a module having various functions for an online interview system.

In particular, the present invention includes a data editing request unit 310 that requests the service server 200 to process moving image data according to needs. When the data editing request unit 310 makes a request that satisfies a specific condition, the text data section of the specific condition section can be found by evaluating the text data and the request condition, and a video that satisfies the request condition in the same manner We could extract the data and provide it by merging it.

The interviewer terminal 300 may be a terminal that is interlocked with a specific server of a company or public institution. If there are a plurality of companies or public institutions, and if the number is singular, the interviewer terminal 300 may have a plurality of forms.

The data editing request unit 310 may be a case in which video data is specifically stored in the service server 200 through the interview terminal 100 and then processing and editing thereof is requested. When a video section including a specific keyword is requested, as described above, the text data is compared with the text data, and then the video data corresponding thereto is extracted.

In the case of setting a video section including a keyword, the selected text data should be selected not in syllables or word units, but in the minimum sentence unit. It would be desirable to expand it a bit further and to have a paragraph-level one. An algorithm that finds this is preferably an algorithm that uses time interval as a function. There may be a certain amount of time interval at the time when thoughts or ideas are transformed, and the selection unit of text data may be determined through an algorithm to determine such time interval. In recent years, since the technology for semantic analysis of text data has been developed, and the technology for extracting text that is related to a part containing a specific keyword has also been developed, this technology may be applied thereto, and this part is within the technical scope of the present invention. It will have to be interpreted.

Meanwhile, questions and other matters are already transmitted to the service server 200 through the interviewer terminal or another server interlocked with the interviewer terminal before the interview of the applicant for the interview through the interview terminal 100 is usually performed. .

Accordingly, the data editing request unit 310 of the interviewer terminal 300 may request a pre-stored questionnaire and a video section satisfying an evaluation item corresponding thereto. In this case, the video data that can be viewed from the interviewer terminal 300 may be video data in the form of extracting only the highlight of the answer of the interviewer. As for the evaluation method of the data determination unit 220 or the extraction method of the data extraction unit 230, those of the aforementioned types may be applied.

It is also possible to separately store the highlight image by providing an algorithm for extracting the highlight image from the service server 200 in advance by changing the implementation method of the platform of the online interview system. The data editing request unit 310 may select corresponding video data through the interviewer terminal 200. If necessary, if there is a need to make a more in-depth selection through the original video data including the idle time, the original video data may be selected. It is also possible to quantify the competency of the interviewer by extracting the summary of the previously stored questionnaire, comparing the text data, and determining the degree of agreement.

This function may be performed through the data determination unit 220 that evaluates the characteristics of text data. Therefore, it can be applied to the form of an automatic interview evaluation system or an automatic headhunting system according to the algorithm of data discrimination, and it can also be applied to the job seeker matching system according to the selection of questions and evaluation items.

The classification system and composition of NCS (National Competency Stanadards) are very well formed. Therefore, if the technical concept of the present invention is applied to an online interview system, an automatic interview evaluation system with improved accuracy and efficiency is used. It seems that implementation is also possible.

Depending on the application type, it can be applied not only for actual employment, but also for interview training. In addition, it may be possible to extract a case including an unsound element such as abusive language through the data determination unit 220 and add it to the evaluation item.

In the present invention, an interview applicant receives question information from a service server and forms video data corresponding to the interview video, receives and stores the video data through a network, and processes and edits the video data through a network. The processed-edited moving picture data is received from the service server, but in the aspect of processing-editing the moving picture data, the detected voice data is converted into text data by detecting the voice data of the moving picture data, and text data is indexed, The video data corresponding to the indexed text data is linked and indexed to match text data and video data, evaluate the characteristics of the text data, select text data according to the evaluation of the data characteristics, and select the selected text It is possible to build a platform that extracts video data corresponding to the index of the data and merges the extracted video data.

100: human resource information processing device, interview terminal
110: signal input unit
120: Talent Information Collection Unit
130: human resource information processing unit
140: talent information storage unit
150: talent information output unit
200: service server
210: data matching unit
220: data discrimination unit
2230: data extraction unit
240: data merging unit
300: interviewer terminal
310: Data editing request unit

Claims (4)

An interview terminal for receiving question information from a service server, forming moving picture data corresponding to the interview image, and transmitting it;
A service server receiving and storing the moving picture data through a network, processing and editing the moving picture data, and transmitting the processed edited moving picture data to an interviewer terminal; And an interviewer terminal receiving the processed and edited video data from the service server through a network,
A signal input unit receiving a signal for requesting talent information including one or more of a keyword related to a person to be discovered and one or more of a research field, a work field, a name, a region, an age, a career, a achievement, a patent, a work place, and a school of the person;
A site that includes at least one of thesis search site, social network service post, blog post, personal dictionary site, newspaper article, social networking site, university site that provides university alumni information, specialized report site for each field, and public patent search site A human resource information collection unit that tracks the past and present information of the human resource from the human resource and receives big data related to the human resource corresponding to the keyword including the relevant data and continuity of the related work;
Related to the talent in a form that can be searched through one or more of the keywords included in the talent information request signal, the research field of the talent, work field, name, region, age, career, achievement, patent, work place, and school A talent information processing unit that converts big data and analyzes the talent-related big data to generate talent status information corresponding to the talent information request signal;
A talent information storage unit that stores the converted big data related to the talent; And
Including a talent information output unit for outputting the talent status information,
The talent information output unit is the current status of the talent according to one or more of the keywords included in the talent information request signal, the research field of the talent, the field of work, name, region, age, career, achievement, patent, work place, and school from school. Visualize information in the form of a Venn diagram to subdivide and output the distribution of manpower.
The step of receiving input from the signal input unit a signal for requesting talent information including keywords related to the talent to be discovered and one or more of the research field, work field, name, region, age, career, achievement, patent, work place, and school of the said talent. ;
A site that includes at least one of thesis search site, social network service post, blog post, personal dictionary site, newspaper article, social networking site, university site providing university alumni information, specialized report site for each field, and open patent search site Searching for, tracking past and present information of the talent, and collecting, at a talent information collection unit, talent-related big data corresponding to keywords containing relevant data and continuity of related work;
It can be searched through one or more of the keyword, the research field of the talent, the field of work, the name, the region, the age, the career, the achievement, the patent, the place of work, and the school from which it is a graduate, included in the request signal for the talent information received from the signal input unit The talent-related big data is converted into a form in which there is a
Storing in an information storage unit; And
Analyzing the talent-related big data stored in the talent information storage unit by the talent information processing unit, generating talent status information corresponding to the talent information request signal, and outputting the generated talent status information from the talent information output unit. Including,
The outputting of the information includes the keyword included in the talent information request signal, the research field of the talent, the field of work, the name, the region, the age, the career, the achievement, the patent, the place of work, Outputs the above talent status information visualized in the form of a Venn diagram that subdivides the distribution of
The service server detects voice data of moving picture data, converts the detected voice data into text data, indexes text data, indexes the moving picture data corresponding to the indexed text data, and indexes the text data. A data matching unit that matches the video data with the video data;
A data determination unit for evaluating the characteristics of the text data;
A data extracting unit for selecting the text data according to the evaluation of the data discriminating unit and extracting moving image data corresponding to the index of the selected text data; And
Further comprising a data merging unit for merging the moving image data obtained through the data extraction unit,
The interviewer terminal includes a data editing request unit for requesting processing and editing of the video data so that the video section including a specific keyword is included in the service server, and the method of indexing the text data in the data matching unit is a function of time. Indexed with, but consists of words in the text,
The data determination unit evaluates idle time by evaluating the interval of time indexed by word unit as a function, and also evaluates whether a specific keyword is included in the text data, and the data extraction unit is text data for extracting video data In selecting, except for the part evaluated as the idle time part of the text data, text data containing a specific keyword is selected in a sentence unit containing a word containing a specific keyword, and the index of the selected text data An intelligent recruitment support platform based on online interview video analysis and social media information analysis that extracts video data corresponding to. The method of claim 1,
The step of converting and storing the talent-related big data is an intelligent recruitment support platform based on online interview video analysis and social media information analysis using one or more of Hadoop, HBASE, and NoSql (Cassandra). The method of claim 1,
An intelligent recruitment support platform based on online interview video analysis and social media information analysis that analyzes the talent-related big data by an ontology-based big data analysis method. The method of claim 1,
The step of outputting the information is a graph based on one or more of the keywords included in the talent information request signal, the research field of the talent, work field, name, region, age, career, achievement, patent, work place, and school An intelligent recruitment support platform based on online interview video analysis and social media information analysis that outputs the above talent status information visualized in the form of.

Images