KR20200129024A - Apparatus and method for recommending job - Google Patents

Abstract

Disclosed are a job recommendation device for recommending a job to a job seeker using a job classification filter generated based on job competency standard classification data, and a job recommendation method and a recording medium thereof. According to an embodiment of the present invention, a job recommendation device comprises: a database for storing a job classification filter generated based on job competency standard classification data including competency unit documents related to competency units; and a job recommendation unit which recommends a job to a job seeker based on competency-related data of the job seeker and the job classification filter. The job classification filter may include at least one of a degree of association between the competency units and a degree of association between the competency units and words extracted from the competency unit documents.

Description

Job recommendation device and job recommendation method {APPARATUS AND METHOD FOR RECOMMENDING JOB}

The present invention relates to a job recommendation device and a job recommendation method, and more particularly, to a job recommendation device and a job recommendation method for recommending a job to a job seeker using a job classification filter generated based on job competency standard classification data. .

Recently, the youth unemployment problem has been recognized as a social issue, and the government/local governments are also seeking solutions to resolve the youth unemployment problem. One of the main causes of the youth unemployment problem is that job seekers do not know exactly what job is right for them. Even if they have succeeded in getting a job, many of the younger employed often leave early because they do not have the aptitude for the job required by the company. Therefore, in order to solve the youth unemployment problem, it is possible to reduce job mismatch between recruiters and job seekers in companies or public institutions, recommend suitable jobs to job seekers, and recommend companies or public institutions that meet the job recommendations of job seekers. A plan is required.

As a conventional technique for recommending jobs to job seekers, there is a method of recommending jobs that match the job seeker's history, such as the job seeker's education, qualifications, education programs completed, and experience through learning based on the employment data of the employed. Exaggerated information may be included in their learning data (for example, resume, self-introduction, portfolio, etc.), so if the accuracy of the learning data is not high, there may be a problem that it is not possible to recommend a suitable job to job seekers. Meanwhile, the government introduced the National Competence Standards (NCS) to check the knowledge, skills, and abilities required for each job in the industrial field to be used as an important standard when hiring talents and to improve job mismatching. I did. However, most job seekers and companies lack awareness of NCS, and are having difficulty finding suitable jobs based on NCS.

The present invention is to provide a job recommendation device, a job recommendation method, and a recording medium for recommending a job to a job seeker by using a job classification filter generated based on job competency standard classification data.

In addition, the present invention is to provide a job recommendation device, a job recommendation method, and a recording medium capable of recommending a job suitable for a job seeker based on a job classification filter prepared with an objective, reasonable, and consistent standard.

The problem to be solved by the present invention is not limited to the problems mentioned above. Other technical problems that are not mentioned will be clearly understood by those of ordinary skill in the art from the following description.

A job recommendation apparatus according to an aspect of the present invention includes: a database for storing a job classification filter generated based on job competency standard classification data including competency unit documents related to competency units; And a job recommendation unit that recommends a job to the job seeker based on the job-seeker's competency-related data and the job classification filter. The job classification filter includes at least one of a degree of association between the competency units and a degree of association between the competency units and words extracted from the competency unit documents.

The job classification filter includes at least one of a first job classification filter and a second job classification filter, and the first job classification filter includes a degree of association between the competency units and words extracted from the competency unit documents. And, the second job classification filter may indicate a degree of correlation between the competency units.

In the first job classification filter, competency units are arranged in a first unit of a row unit and a column unit, words extracted from the competency unit documents are arranged in the other second unit, and the first A two-dimensional matrix having a degree of association between the capability unit document associated with the capability unit of the unit and the word of the second unit as each component value may be included.

In an embodiment, the job recommendation unit may classify or recommend a job matching the job-seeker's competency-related data based on the job-seeker's competency-related data and the first job classification filter.

The job recommendation unit may include a word extracting unit for extracting competency related words corresponding to words arranged in the second unit of the first job classification filter from among words of the competency related data; A vectorization unit for generating a feature vector by vectorizing the competency-related words according to an arrangement order of words arranged in the second unit of the first job classification filter; A similarity calculation unit for calculating a similarity between the feature vector and a competency unit vector corresponding to each competency unit of the first job classification filter; And a recommended job determination unit for determining a recommended job in the order of a high degree of similarity between the capability unit vectors corresponding to the capability units and the feature vector.

The job recommendation unit may classify jobs related to recruitment based on recruitment-related data and the first job classification filter, and recommend one or more employment information to the job seeker based on the job classification information of the recruiter. The job recommendation unit may include a word extracting unit for extracting employment-related words corresponding to words arranged in the second unit of the first job classification filter from among words of the employment-related data; A vectorization unit that vectorizes the recruitment-related words according to an arrangement order of words arranged in the second unit of the first job classification filter to generate a feature vector; A similarity calculation unit for calculating a similarity between the feature vector and a competency unit vector corresponding to each competency unit of the first job classification filter; And a job classification unit that determines one or more competency units related to the employer in the order of similarity between the competency unit vectors corresponding to the competency units and the feature vector.

In the second job classification filter, competency units are arranged in a first unit of a row unit and a column unit, the competency units are arranged in the other second unit, and are associated with the competency units of the first unit. A two-dimensional matrix having a relationship between the first capability unit document and the second capability unit document associated with the capability unit of the second unit as each component value may be included.

In an embodiment, the job recommendation unit may classify or recommend a job matching the job-seeker's competency-related data based on the job-seeker's competency-related data and the second job classification filter.

The job recommendation unit may include a word extracting unit that extracts words related to the competency unit of the job seeker from the competency related data; A job competency analysis unit that determines one or more first competency units related to the job seeker from words extracted from the competency-related data; And a recommended job determination unit that determines at least one second competency unit based on the similarity with the first competency unit based on the second job classification filter and determines the second competency unit as a recommended job. The recommended job determination unit may apply the job seeker's competency-related history information related to the first competency unit to the second job classification filter to reflect the competency-related history information for other competency units.

The job recommendation unit may recommend one or more recruitment information to the job seeker by analyzing the job competency criteria related to recruitment based on the recruitment-related data and the second job classification filter. The job recommendation unit may include a word extracting unit that extracts words related to competency units from the recruitment-related data; And at least one first competency unit related to the recruiter is determined based on words extracted from the recruitment-related data, and at least one second competency in the order of similarity with the first competency unit based on the second job classification filter. It may include a job competency criteria analysis unit that extracts units.

The apparatus for recommending a job according to an embodiment of the present invention may further include a job classification filter generator that generates the first job classification filter based on the job capability standard classification data. The job classification filter generation unit may include at least one of a first job classification filter generation unit that generates the first job classification filter and a second job classification filter generation unit that generates the second job classification filter.

The first job classification filter generator may generate the first job classification filter by calculating a degree of association between the competency unit documents and words extracted from the competency unit documents. The second job classification filter generator may generate the second job classification filter by calculating a relationship between the competency unit documents based on words extracted from the competency unit documents.

The first job classification filter generation unit may include a document extracting unit for extracting NCS competency unit documents for each major classification among National Competency Standards (NCS) classification data; A word extracting unit for extracting words from all of the NCS capability unit documents belonging to the same major classification; Association between NCS capability units and the extracted words based on the word frequency inverse document frequency (TF-IDF) between words extracted from the NCS capability unit documents and the NCS capability unit documents A correlation analysis unit for analyzing a degree; And a normalization unit for normalizing by dividing a 2D matrix having a degree of association between the NCS capability units analyzed by the association analysis unit and the extracted words as component values, by a set value. The document extracting unit may extract a document including the name, definition, job basic ability, evaluation items, performance criteria for each competency unit element, knowledge, skills, and attitude items of the NCS competency unit.

The second job classification filter generation unit may include a document extracting unit for extracting NCS competency unit documents from National Competency Standards (NCS) classification data; A word extracting unit that extracts words from all of the NCS capability unit documents; And an inter-class relationship analysis unit that analyzes a degree of association between NCS capability unit documents based on the similarity of words extracted from the NCS capability unit documents.

The job recommendation unit may display the recommended job in a different display method depending on whether the job recommendation determined by the job recommendation unit matches the job selected by the job seeker.

The competency-related data may include at least one of portfolio, resume, self-introduction, SNS data, and data collected by a chatbot collecting history information in the form of a conversation with the job seeker.

According to another aspect of the present invention, a job recommendation method is provided, including, by a job recommendation unit, recommending a job to the job seeker based on data related to the job seeker's competency and a job classification filter. The job classification filter is generated based on competency unit documents related to competency units of the job competency standard classification data, the degree of relationship between the competency units, and words extracted from the competency units and the competency unit documents It may include at least one of the degree of association between.

The recommending of the job may include: extracting, by a word extracting unit, competency related words corresponding to words arranged in the second unit of the job classification filter from among words of the competency related data; Generating a feature vector by vectorizing the competency-related words according to an arrangement order of words arranged in the second unit of the job classification filter by a vectorization unit; Calculating a similarity between the feature vector and the capability unit vector corresponding to each capability unit of the job classification filter, by a similarity calculation unit; And determining, by the recommended job determination unit, a job recommendation in order of a high degree of similarity between the competency unit vectors corresponding to the competency units and the feature vector.

The step of recommending the job may include extracting, by a word extracting unit, words related to the competency unit of the job seeker from the competency related data; Determining one or more first competency units related to the job seeker from words extracted from the competency-related data, by a job competency analysis unit; And determining at least one second competency unit based on the similarity with the first competency unit based on the job classification filter, and determining the second competency unit as a recommended job. can do.

The job recommendation method according to an embodiment of the present invention may further include generating the job classification filter based on the job capability standard classification data. The generating of the job classification filter may include: extracting NCS competency unit documents for each major classification among National Competency Standards (NCS) classification data by a document extraction unit; Extracting, by a word extracting unit, words from all of the NCS capability unit documents belonging to the same major classification; By the association analysis unit, based on the word frequency inverse document frequency (TF-IDF; Term Frequency Inverse Document Frequency) between the words extracted from the NCS capability unit documents and the NCS capability unit documents, the NCS capability units and Analyzing a degree of association between the extracted words; And a 2D matrix having a degree of association between the NCS capability units analyzed by the association analysis unit and the extracted words as component values, by a normalization unit, so that the size of the 2D matrix becomes a set value. It may include the step of.

The generating the job classification filter may include: extracting NCS competency unit documents from National Competency Standards (NCS) classification data by a document extracting unit; Extracting, by a word extracting unit, words from all of the NCS capability unit documents; And analyzing a correlation between the NCS capability unit documents based on the similarity of words extracted from the NCS capability unit documents, by the relationship analysis unit between classifications.

In the job recommendation method according to an embodiment of the present invention, the job classification filter update unit collects update information of the job competency standard classification data, and sets the job classification filter based on the update information of the job competency standard classification data. It may further include the step of updating.

According to another aspect of the present invention, a computer-readable recording medium in which a program for executing the job recommendation method is recorded is provided.

According to an embodiment of the present invention, a job recommendation device, a job recommendation method, and a recording medium for recommending a job to a job seeker using a job classification filter generated based on job competency standard classification data are provided.

In addition, according to an embodiment of the present invention, it is possible to recommend a job suitable for a job seeker based on a job classification filter prepared with an objective, reasonable and consistent standard, and job mismatch between the job seeker and the recruiter can be resolved.

The effect of the present invention is not limited to the above-described effects. Effects not mentioned will be clearly understood by those of ordinary skill in the art from the present specification and the accompanying drawings.

1 is a configuration diagram of a job recommendation device according to an embodiment of the present invention.
2 is a block diagram of a job recommendation server constituting a job recommendation device according to an embodiment of the present invention.
3 is a configuration diagram of a job classification filter generating unit constituting a job recommendation device according to an embodiment of the present invention.
4 is a block diagram of a job recommendation unit constituting a job recommendation device according to an embodiment of the present invention.
5 is a configuration diagram of a first job classification filter generating unit constituting a job recommendation device according to an embodiment of the present invention.
6A and 6B are diagrams illustrating competency unit documents of job competency standard classification data.
7 is an exemplary diagram of a first job classification filter according to an embodiment of the present invention.
8 is a configuration diagram of a first job recommendation unit constituting a job recommendation device according to an embodiment of the present invention.
9 is an exemplary diagram of jobs recommended to job seekers according to an embodiment of the present invention.
10 is a configuration diagram of a first job recommendation unit constituting a job recommendation device according to another embodiment of the present invention.
11 is a configuration diagram of a second job classification filter generating unit constituting a job recommendation device according to an embodiment of the present invention.
12 is an exemplary diagram of a second job classification filter according to an embodiment of the present invention.
13 is a configuration diagram of a second job recommendation unit constituting a job recommendation device according to an embodiment of the present invention.
14 is a block diagram of a second job recommendation unit constituting a job recommendation device according to another embodiment of the present invention.

Other advantages and features of the present invention, and a method of achieving them will become apparent with reference to embodiments to be described later in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, and the present invention is only defined by the scope of the claims. Even if not defined, all terms (including technical or scientific terms) used herein have the same meaning as commonly accepted by universal technology in the prior art to which this invention belongs. General descriptions of known configurations may be omitted so as not to obscure the subject matter of the present invention. In the drawings of the present invention, the same reference numerals are used as much as possible for the same or corresponding configurations.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as "comprise", "have" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or a combination thereof described in the specification. It is to be understood that the possibility of the presence or addition of other features, numbers, steps, actions, components, parts, or combinations thereof, or any further features, is not excluded in advance.

The'~ unit' used throughout this specification is a unit that processes at least one function or operation, and may mean, for example, a hardware component such as software, FPGA, or ASIC. However,'~ part' is not limited to software or hardware. The'~ unit' may be configured to be in an addressable storage medium, or may be configured to reproduce one or more processors. As an example,'~ unit' refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, and subs. Routines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The components and functions provided by the'~ unit' may be performed separately by a plurality of elements and the'~ units', or may be integrated with other additional elements.

1 is a configuration diagram of a job recommendation device according to an embodiment of the present invention. Referring to FIG. 1, a job recommendation device 10 according to an embodiment of the present invention may include a job recommendation server 30 and a database 40. The job recommendation server 30 may collect competency-related data from the job seeker terminal 20 and recommend a job suitable for the job seeker using a job classification filter generated based on job competency standard classification data. The job seeker terminal 20 is a terminal used by job seekers, and may be implemented as, for example, a smartphone, a tablet PC, a video phone, a desktop PC, a laptop PC, a netbook computer, a workstation, a PDA, a PMP, a wearable device, etc. It is not limited thereto. The job classification filter according to an embodiment of the present invention will be described in more detail later with reference to the drawings of FIGS. 7 and 12.

The database 40 includes data related to competency of job seekers, data related to recruitment of recruiters such as companies or public institutions, job competency standard classification data such as NCS data, refined NCS data, and a job for recommending jobs suitable for job seekers. You can store information such as classification filters. The database 40 is not necessarily interpreted as being separated from the job recommendation server 30 and provided separately, and should be understood as a concept including a memory provided in the job recommendation server 30. The job recommendation server 30 may access the job seeker terminal, the recruiter terminal, and the database 40 through various types of communication. Here, various types of communication may be implemented by a network interface capable of wired and/or wireless communication, for example, short-range communication (eg, Bluetooth, WiFi direct, or IrDA (infrared data association)) or long-distance communication (eg, cellular It may be a network, the Internet, or a computer network (eg, LAN or WAN), but is not limited thereto. Wired/wireless network interfaces communicate according to various communication standards such as IEEE, zigbee, 3rd Generation (3G), 3rd Generation Partnership Project (3GPP), Long Term Evoloution (LTE), 5th Generation (5G), etc. It may be configured as a network interface that performs, but is not limited thereto.

2 is a block diagram of a job recommendation server constituting a job recommendation device according to an embodiment of the present invention. 1 and 2, the job recommendation server 30 may include a job classification filter generator 100, a job classification filter 200, and a job recommendation unit 300. The job classification filter generator 100 may generate one or a plurality of different job classification filters 200 based on job competency standard classification data. In an embodiment, in order to recommend a job suitable for a job seeker on an objective basis, the job classification filter 200 does not use the job-seeker's competency-related data or the recruiter's employment-related data, but only based on the information of the job competency standard classification data itself. Can be calculated as The job classification filter 200 generated by the job classification filter generation unit 100 may be stored in the database 40. The job recommendation unit 300 may recommend a job suitable for the job seeker based on job-seeker-related data and the job classification filter 200 generated by the job classification filter generation unit 100.

In an embodiment, the job competency standard classification data may include a National Competency Standards (NCS) classification table and NCS competency unit documents. The NCS classification table is a systematized content of knowledge, skills, and abilities required to perform a job in the industrial field.Large category (24) → Medium category (81) → Small category (256) → Subclass (1001) → It has a classification system in the order of competency units (12,120 units). Each competency unit is subdivided into'capability unit factor','performance criteria', and'knowledge/skill/attitude' and is defined.

In an embodiment, the job classification filter generation unit 100 extracts NCS competency unit documents, extracts words from the extracted competency unit documents, and then relates the degree of relationship between competency unit documents and words or between competency unit documents. Job classification filters can be created by calculating degrees. Each of the 12,120 NCS competency unit documents contains the'Competency Unit Classification Number','Name','Definition','Evaluation Guidelines','Basic Vocational Competency','Evaluation Items', and competency unit elements belonging to the competency unit. It may be a document including information such as'name','conformity','knowledge','technology', and'attitude'.

3 is a configuration diagram of a job classification filter generating unit constituting a job recommendation device according to an embodiment of the present invention. 4 is a block diagram of a job recommendation unit constituting a job recommendation device according to an embodiment of the present invention. 3 and 4, the job classification filter generation unit 100 includes a first job classification filter generation unit 120, a second job classification filter generation unit 140, and a job classification filter update unit 160. Can include. The job classification filter 200 may include a first job classification filter 220 and a second job classification filter 240. The job recommendation unit 300 may include a first job recommendation unit 320 and a second job recommendation unit 340.

The first job classification filter generation unit 120 may generate a first job classification filter 220 of a two-dimensional matrix representing a relationship diagram between competency unit documents and words extracted from all of the competency unit documents. The second job classification filter generator 120 may generate a second job classification filter 240 of a two-dimensional matrix indicating a degree of association between competency unit documents. The first job recommendation unit 320 may recommend a job to a job seeker based on the first job classification filter 220 generated by the first job classification filter generation unit 120. The second job recommendation unit 340 may recommend a job to a job seeker based on the second job classification filter 240 generated by the second job classification filter generation unit 140. Hereinafter, the first job classification filter 220 will be described first, and then the second job classification filter 240 will be described.

5 is a configuration diagram of a first job classification filter generating unit constituting a job recommendation device according to an embodiment of the present invention. 3 and 5, the first job classification filter generation unit 120 includes a document extraction unit 122, a word extraction unit 126, a correlation analysis unit 130, and a normalization unit 132. It may include. The document extractor 122 may extract competency unit documents 124 from the job competency standard classification data 400. In an embodiment, the document extractor 122 may extract capability unit documents 124 for each of 24 NCS major categories. The number n of the competency unit documents 124 extracted for each major category may be different, and n represents the number of competency units belonging to the major category. The n number of competency unit documents 124 belonging to the same major category can be used to calculate the first job classification filter for the corresponding major category, and as many as the number of major categories (for example, 24) first job classification filters Can be calculated.

6A and 6B are diagrams illustrating competency unit documents of job competency standard classification data. The document extracting unit 122 is configured for each competency unit:'Competency Unit Classification Number','Name','Level','Definition','Basic Vocational Competency','Evaluation Items','Performance Standard','Performance Unit Element' A document 124 including information such as knowledge','skill', and'attitude' may be extracted. The word extraction unit 126 may extract words 128 from all of the n capability unit documents 124. The word extracting unit 126 may include a morpheme analyzer that divides sentences in the competency unit documents 124 into words in morpheme units and tagging parts of speech. In an embodiment, the word extracting unit 126 extracts specific morphemes and morpheme combinations such as nouns and roots after morpheme analysis, processes a self-made stopword list, and then applies N-grams such as Bigram in the case of continuous nouns can do.

The correlation analysis unit 130 may analyze a degree of association between competency unit documents and words extracted from competency unit documents. In an embodiment, the association analysis unit 130 includes n capability unit documents 124 and m words extracted from n capability unit documents 124 based on TF-IDF (Term Frequency Inverse Document Frequency). The degree of association between them can be analyzed. The degree of association between the competency unit documents 124 and the words 128 may be calculated according to Equations 1 to 3 below, or may be calculated according to Equations 1, 2, and 4 below.

[Equation 1]

[Equation 2]

[Equation 3]

[Equation 4]

In Equations 1 to 4,'tfidf(t,d,D)' is the word frequency inverse document frequency,'t' is a word,'d' is a document, D is a set of all documents, and tf(t,d) is a word Frequency, idf(t,D) is the reverse document frequency,'f _t,d ' is the frequency of the word t in the document d, and'df(t,d)' is the number of documents containing the word t Indicate the number.

The association analysis unit 130 applies not only TF-IDF, but also non-negative matrix factorization (NMF), word embedding technology such as Word2Vec/Doc2Vec, and Also, it is possible to analyze the degree of association between words extracted from competency unit documents.

When the number of words 128 extracted from the n capability unit documents 124 is m, the degree of association between the capability unit documents 124 and the words 128 can be represented by an n×m matrix. have. The normalization unit 132 may generate a normalized matrix by dividing each component of the n×m matrix by a set value so that the size of the matrix of size n×m is normalized to 1. The set value may be different depending on each component of the n×m matrix. Also, the set value may be different depending on the normalization method. On the other hand, in the present embodiment, the matrix is normalized to have a size of 1, but it may be modified to an appropriate size such as 10 or 100. Accordingly, a normalized first job classification filter 220 may be generated for each major classification. By normalizing the first job classification filter 220, it is possible to prevent the priority of recommended jobs belonging to different major categories from being influenced by different matrix sizes between major categories, and by competency units within a specific major category (large category with a large matrix size). It is possible to prevent concentration of recommended jobs, and recommend jobs (competency units) suitable for job seekers in large categories with large matrix sizes as well as large categories with small matrix sizes.

7 is an exemplary diagram of a first job classification filter according to an embodiment of the present invention. 5 and 7, the first job classification filter 220 includes competency units (a ₁ , a ₂ , ..., a _n-1 , a _n ) in a row unit (first unit). ) Are arranged, and words 128 extracted from competency unit documents 124 in a column unit (second unit) (v ₁ , v ₂ , ..., v _m-1 , v _m ) This can be expressed as an arrayed two-dimensional matrix. Each component value of the 2D matrix of the first job classification filter 220 may be a value obtained by normalizing the TF-IDF value of the word so that the matrix size becomes a set value. For example, it can be normalized by dividing the set value so that the size of the matrix becomes 1. The first job classification filter 220 may be regarded as a set of row vectors having a size of 1×m for each of the n job unit capabilities. The number m of words 128 may be approximately tens of thousands to hundreds of thousands of words.

In the above, for an embodiment of the first job classification filter 220 in which competency units are sequentially arranged in row units (first unit) and words extracted from competency unit documents are sequentially arranged in column units (second unit). Although described, in the first job classification filter 220, competency units may be arranged in column units, and words extracted from competency unit documents may be arranged in row units. In this case, each capability unit vector may be provided as a column vector.

Referring back to FIG. 3, the job classification filter update unit 160 collects update information of job competency standard classification data through periodic crawling technology, and classifies job based on the update information of job competency standard classification data. Filters can be updated. Job competency standard classification data can be changed or updated to reflect future social changes and industrial demands, and can be added or deleted if there is a job that is newly created or eliminated. The job classification filter update unit 160 may update the job classification filter when the job competency standard classification data is updated.

For example, the job classification filter update unit 160 finds a new competency unit classification number or a competency unit having a new name in the job competency standard classification data collected by crawling, or the competency unit classification numbers or names of existing competency units. Among them, if the deleted competency unit classification number or name exists, or by a crawling technique, the'vocational basic competency','evaluation items','performance criteria','knowledge', and When it is determined that information such as'technology' and'attitude' has changed, it is determined that the job competency standard classification data has been changed or updated, and the job classification filter can be updated.

8 is a configuration diagram of a first job recommendation unit constituting a job recommendation device according to an embodiment of the present invention. 4, 7 and 8, the first job recommendation unit 320 may recommend one or more jobs to the job seeker based on the job seeker's competency-related data 500 and the first job classification filter 220. have. In an embodiment, the competency-related data 500 may be job search-related data. In an embodiment, the competency-related data 500 includes data collected by a chatbot that collects education information, holding certificate information, portfolio, resume, self-introduction, SNS data, and history information in the form of conversation with the job seeker, and the job seeker's voice. It may include digitized personal information such as data collected through data, user's social relationship information, SNS interests, content viewing records, search information, and the like. The competency-related data 500 may be collected from a job-seeker terminal, or may be collected from a job-seeker's SNS or email, a job-seeking relay server, a school or a state administrative agency server that stores and manages information on job seekers. The collected information may be digitized personal information such as education information, social relations information of job seekers, interests of SNS, content viewing records, and search information. The first job recommendation unit 320 may include a word extraction unit 322, a vectorization unit 324, a similarity calculation unit 326, and a recommended job determination unit 328.

The word extraction unit 322 may extract competency-related words from the competency-related data 500 of the job seeker. The word extracting unit 322 may include a morpheme analyzer for dividing sentences in the competency-related data 500 into words in morpheme units and tagging parts of speech. In an embodiment, the word extracting unit 322 extracts specific morphemes and morpheme combinations such as nouns and roots after morpheme analysis, processes a self-made stopword list, and then applies N-grams such as Bigram in the case of noun continuation. can do. The word extractor 322 may extract only words corresponding to the words 128 defined in the first job classification filter 220 from among words of the competency-related data 500.

The vectorization unit 324 may vectorize words extracted from the competency-related data 500 by the word extraction unit 322 according to the arrangement order of the words 128 defined in the first job classification filter 220. . In the first job classification filter 220, the order of words 128 is arranged in column units. The vectorization unit 324 extracts k (k is an integer of 0 or more and m or less) from the competency-related data 500 according to the arrangement order of words 128 defined in the first job classification filter 220 Can be converted into a 1×m row vector. In this case, in the row vector transformed from words extracted from the capability-related data 500, k components may have a value greater than 0, and the remaining (m-k) components may have a value of 0. The component value of each of the k words may be determined as a value corresponding to the Term Frequency (TF) of the word in the competency-related data 500, or a value irrespective of the frequency by applying a binary score. It may be decided. When a word frequency (TF) is used, a normalized term frequency can be used to calculate the similarity (eg, cosine similarity). In the case of using a binary score, it can be applied by taking the square and then dot product.

For example, assuming that the word'architectural design' is included twice in the competency-related data 500, and the word corresponding to'architectural design' in the first job classification filter 220 is defined in the 12145th column, When the frequency of the word is not considered, the value of the 12145th component of the primary vector generated by the vectorization unit 324 is determined as '1', and when the frequency of the word is considered, 1 generated by the vectorization unit 324 The value of the 12145th component of the difference vector may be determined as '2'. Then, a normalized row vector (feature vector) can be generated by a vector normalization operation. The similarity calculation unit 326 is based on the row vector (feature vector) generated from the competency-related data 500 by the vectorization unit 324 and the feature vector and competency units of the job seeker based on the first job classification filter 220 The degree of similarity between the livers can be calculated. In an embodiment, the similarity calculation unit 326 may calculate a similarity (eg, cosine similarity) between a feature vector of a job seeker and a row vector (capability unit vector) of each of the competency units.

In an embodiment, the cosine similarity may be calculated as a value obtained by dividing the feature vector of the job seeker and the row vector of each of the m capability units by the size of the feature vector and the size of the row vector of the capability unit. When the feature vector is normalized, the operation of dividing by the size of the feature vector may be omitted. The similarity calculation unit 326 may calculate a similarity with a feature vector of a job seeker for each of n competency units belonging to one major category. The recommended job determination unit 328 may determine the recommended job to be recommended to the job seeker by determining the priority of the competency units in the order of the highest similarity to the job seeker's characteristic vector among n similarity values between the job seeker's feature vector and the competency units. have.

In an embodiment, the recommended job determination unit 328 may determine a recommended job from one or a plurality of major categories in an order of high similarity between the job seeker's feature vector and competency units for each major category. Alternatively, when a job seeker enters an input to select a specific major category using a user interface, jobs (competency units) belonging to the major category may be recommended based on the job classification filter for only the selected major category. In the above, the case of recommending jobs by creating various job classification filters for each major category has been described as an example, and it is also possible to create a job classification filter for all competency unit documents belonging to all major categories to recommend jobs.

According to an embodiment of the present invention, a job classification filter is not created by learning using careers (past job experience), such as a job seeker's resume, portfolio, and self-introduction, but based on only job competency standard classification data. By creating a classification filter and recommending jobs to job seekers, it is possible to prevent error factors from occurring in the job classification filter due to uncertain learning data, and it is suitable for job seekers based on the job classification filter prepared with objective, reasonable, and consistent standards. Can recommend a job.

In the above, a job recommendation device and a job recommendation method according to an embodiment of the present invention have been described using the NCS as an example, but the job recommendation device and the job recommendation method according to the embodiment of the present invention are not only the NCS but also the National Occupational Standards (NOS). , VAAS (Vocational Ability Assessment Standards), NSS (National Skill Standard), NQF (National Qualification Framework), KQF (Korean Qualifications Framework), or other job competency standards not listed. In addition, even if it is not defined as a standard, it can be applied to the case of defining knowledge, skills, attitudes, etc. for job competency.

9 is an exemplary diagram of jobs recommended to job seekers according to an embodiment of the present invention. When the job to be recommended to the job seeker is determined, the job recommendation server can visually display the job recommended to the job seeker in the order of major category-medium category-sub category-sub category-competency unit. In addition, when the job seeker selects the desired job using the user interface, different display methods (color, font size, font, shading, font thickness, highlighting, etc.) Jobs selected by the job seeker and recommended jobs determined by the job recommendation server may be displayed.

In the embodiment of FIG. 9, when the job recommended by the job recommendation server matches the job searcher's selected job, the recommended job is displayed in red, and when the job recommended by the job recommendation server is a job not selected by the job seeker If the recommended job is displayed in black and the job selected by the job seeker is not recommended by the job recommendation server, the job selected by the job seeker is indicated in blue, but other various display methods indicate whether the recommended job matches the selected job. It is also possible to indicate.

10 is a configuration diagram of a first job recommendation unit constituting a job recommendation device according to another embodiment of the present invention. 7 and 10, the first job recommendation unit 320 classifies jobs related to recruitment based on the recruitment-related data 600 and the first job classification filter 220, and provides job classification information of the recruiter. Based on it, you can recommend one or more jobs to job seekers. In an embodiment, the recruitment-related data 600 may include recruitment information such as career, educational program, academic background, and job experience required by a recruiter such as a company or a public institution. The recruitment-related data 600 may be collected from a recruiter terminal, or may be collected from a recruiter's SNS or email, a job posting relay server, a school or a national administrative agency server that stores and manages information on the recruiter. In addition, the recruitment-related data 600 may be collected by a chatbot that collects information in the form of a conversation with the recruiter, or may be collected through voice analysis of the recruiter. The first job recommendation unit 320 may include a word extracting unit 330, a vectorization unit 332, a similarity calculation unit 334, and a job classification unit 336.

The word extractor 330 may extract employment-related words from the recruitment-related data 600 of the recruiter. The word extracting unit 330 may include a morpheme analyzer for dividing sentences in the recruitment-related data 600 into words in morpheme units and tagging parts of speech. In an embodiment, the word extracting unit 330 extracts specific morphemes and morpheme combinations such as nouns and roots after morpheme analysis from sentences, processes the self-made stopword list, and in the case of noun continuation, N -Gram can be applied. The word extractor 330 may extract only words corresponding to the words 128 defined in the first job classification filter 220 from among words of the recruitment-related data 600.

The vectorization unit 332 may vectorize words extracted from the recruitment-related data 600 by the word extraction unit 330 according to the arrangement order of the words 128 defined in the first job classification filter 220. . The vectorization unit 332 extracts p (p is an integer of 0 or more and m or less) extracted from the recruitment-related data 600 according to the arrangement order of the words 128 defined in the first job classification filter 220 Can be converted into a 1×m row vector. In this case, in the row vector (feature vector) converted from words extracted from the recruitment-related data 600, p components among m components have a size greater than 0, and the remaining (mp) components have a value of 0. Can have The component value of each of the p words may be determined as a value corresponding to the Term Frequency (TF) of the word in the hiring-related data 600 or may be determined as a value irrespective of the frequency based on the binary score. When a word frequency (TF) is used, a normalized term frequency can be used to calculate the similarity (eg, cosine similarity). In the case of using a binary score, it can be applied by taking the square and then dot product.

When input data (eg, job-seeker's competency data) is input, the binary score is vectorized in consideration of the frequency or vectorized only by the appearance of words without considering the frequency. If it is vectorized by considering the frequency, it is normalized and the score is calculated with cosine similarity. If the frequency is not considered, the normalized or unnormalized vector is squared and then the score is calculated by dot product.

For example, assuming that the word'architectural design' is included three times in the recruitment-related data, and the word corresponding to'architectural design' in the first job classification filter 220 is defined in the 12145th column, the frequency of the word If not considered, the 12145th component value of the primary vector generated by the vectorization unit 332 is determined as '1', and when the frequency of the word is considered, the primary vector generated by the vectorization unit 332 is The 12145th component value may be determined as '3'. Then, normalized row vectors (feature vectors) can be generated by vector normalization.

The similarity calculation unit 334 is based on the row vector (feature vector) generated from the recruitment-related data 600 by the vectorization unit 332 and the feature vector and competency units of the recruiter based on the first job classification filter 220 The degree of similarity between the livers can be calculated. In an embodiment, the similarity calculation unit 334 may calculate a similarity (eg, cosine similarity) between a feature vector of the employer and a row vector of each of the capability units. In an embodiment, the cosine similarity may be calculated as a value obtained by dividing the adopter's feature vector and the row vector of each of the m capability units by the size of the adopter's feature vector and the row vector size of the capability unit. The similarity calculation unit 334 may calculate a similarity with a feature vector of an adopter for each of n capability units belonging to one major classification. The job classification unit 336 determines the priorities of the competency units in the order of the highest similarity to the recruiter's characteristic vector among n similarity values between the recruiter's characteristic vector and the competency units, and assigns the job (capacity unit) related to the recruiter. It can be classified according to competency standard classification.

In an embodiment, the job classification unit 336 may provide job classification results from one or more major categories, in an order of high similarity between the feature vectors of the recruiter and the competency units, for all major categories. Alternatively, when the recruiter enters an input for selecting a specific major category using the user interface, jobs belonging to the major category may be classified based on the job classification filter for only the selected major category. In the above, the case of classifying the job of the recruiter by creating various job classification filters for each major category has been described as an example.It is also possible to classify jobs related to recruiters by creating one job classification filter for all competency unit documents belonging to all major categories. It is possible.

When one or more job classification information related to the recruiter is determined, recruitment information is provided to the job seeker matching the job classification information of the corresponding recruiter, and job matching between the job seeker and the recruiter can be made based on the competency unit. Therefore, recruiters can be recommended for talents that meet the job aptitudes of companies and public institutions, and recruiters are provided with recruitment information that can perform appropriate jobs based on competency units, thereby preventing job mismatch between recruiters and recruiters. Can be prevented.

Hereinafter, an embodiment in which the second job classification filter 240 is generated to recommend jobs will be described. 11 is a configuration diagram of a second job classification filter generating unit constituting a job recommendation device according to an embodiment of the present invention. 3 and 11, the second job classification filter generation unit 140 is a second job classification filter 240 of a two-dimensional matrix indicating the degree of association between documents 144 extracted from job competency standard classification data. Can be created.

The second job classification filter generation unit 140 may include a document extracting unit 142, a word extracting unit 146, and a relationship analysis unit 148 between classifications. The document extraction unit 142 may extract competency unit documents 144 from the job competency standard classification data 400. In an embodiment, the document extracting unit 142 may extract capability unit documents 144 for each of 24 NCS major categories. The number n of the competency unit documents 144 extracted for each major category may be different, and n represents the number of competency units belonging to the major category. The n number of competency unit documents 144 belonging to the same major category can be used to calculate a second job classification filter for the corresponding major category, and as many as the number of major categories (for example, 24) second job classification filters (240) can be calculated.

The word extractor 146 may extract words from all of the n capability unit documents 144. The word extracting unit 146 may include a morpheme analyzer for dividing sentences in the competency unit documents 144 into words in morpheme units and tagging parts of speech. In an embodiment, the word extracting unit 146 extracts specific morphemes and morpheme combinations such as nouns and roots after morpheme analysis, processes the self-made stopword list, and then applies N-grams such as Bigram in the case of noun continuation. can do.

The relationship between classifications analysis unit 148 may analyze the degree of correlation between the competency unit documents 144 based on the similarity of words extracted from the competency unit documents 144. In an embodiment, the degree of similarity between words may be expressed as a degree of similarity of capability unit document vectors having a degree of association between extracted words as a component. In an embodiment, the relationship analysis unit 148 between classifications may analyze a degree of correlation between n capability unit documents 144 based on a TF-IDF (Term Frequency Inverse Document Frequency).

In an embodiment, when m words are extracted from all competency unit documents 144 by the word extracting unit 146, the interclassification relationship analysis unit 148 maps the order of all words to the column order. You can create a row vector of size ×m. The relationship between classification analysis unit 148 generates a row vector for each competency unit document 144 from a set of words extracted from each competency unit document 144, and the number of competency unit documents 144 ( n) number of row vectors can be created. Thereafter, for all n×n combinations of the n capability unit documents 144, for example, the cosine similarity between the row vectors of the two capability unit documents 144 according to Equations 1 to 4 is calculated. By calculating, the relationship between the competency unit documents 144 may be analyzed. Relationship values between the competency unit documents 144 may be represented by a two-dimensional matrix having an n×n size, and the two-dimensional matrix may be stored as the second job classification filter 240.

12 is an exemplary diagram of a second job classification filter according to an embodiment of the present invention. 11 and 12, the second job classification filter 240 includes competency units (a ₁ , a ₂ , ..., a _n-1 , a _n ) in a row unit (first unit). ) Is arranged, and the capability units (a ₁ , a ₂ , ..., a _n-1 , a _n ) are arranged in a column unit (a second unit). Capability units (a ₁ , a ₂ , ..., a _n-1 , a _n ) may be defined in the same order of arrangement in row and column units. Each component value of the 2D matrix of the second job classification filter 240 may be a similarity value between two row vectors generated from two word sets each extracted from two competency unit documents. Since the diagonal component of the second job classification filter 240 is a similarity value between two identical row vectors extracted from the same competency unit document, the component value is 1. The second job classification filter 240 may be regarded as a set of row vectors having a size of 1×n for each of the n job unit capability units. A large component value in the second job classification filter 240 means that the similarity between the two corresponding competency units is large, and in the opposite case, the similarity between the two competency units is small.

13 is a configuration diagram of a second job recommendation unit constituting a job recommendation device according to an embodiment of the present invention. 4, 12, and 13, the second job recommendation unit 340 may recommend one or more jobs to the job seeker based on the job seeker's competency-related data 500 and the second job classification filter 240. have. The second job recommendation unit 340 may include a word extracting unit 342, a job capability analysis unit 344, and a recommended job determination unit 346.

The word extracting unit 342 may extract competency-related words from the competency-related data 500 of the job seeker. The word extracting unit 342 may include a morpheme analyzer for dividing sentences in the competency-related data 500 into words in morpheme units and tagging parts of speech. In an embodiment, the word extracting unit 342 extracts specific morphemes and morpheme combinations such as nouns and roots after morpheme analysis, processes a self-made stopword list, and then applies N-grams such as Bigram in the case of continuous nouns And, you can extract words related to the job seeker's competency unit.

The job competency analysis unit 344 determines one or more first competency units related to the job seeker from words extracted from the competency-related data 500, and determines the first competency unit and the high level based on the second job classification filter 240. One or more second units of capability having similarity may be extracted. As another example, when the level score for the first competency unit is input by the job seeker, or when completion information such as a career or education program related to the first competency unit is input, the job competency analysis unit 344 is a second job classification filter. Based on the similarity matrix relationship between competency units of (240), the level score or completion information for the first competency unit may be reflected as the level score or completion information for other competency units.

For example, in the second job classification filter 240 shown in FIG. 12, the a _n- th competency unit (the second competency unit) has a relevance of 0.30 to the a _n- 1th competency unit (first competency unit). , a If the job seeker's level score related to the _n- 1th competency unit is entered as 90 points, measured/evaluated or analyzed, the job seeker's level score can be reflected as 27 points for the _nth competency unit as well. Accordingly, based on the second job classification filter 240, the competency-related history for the first competency unit of the job seeker can be expanded to other competency units, so that various competency units suitable for the job seeker can be recommended to the job seeker.

In an embodiment, the recommended job determination unit 346 may determine the recommended job from one or more major categories by analyzing competency units related to the competency-related data 500 for each major category. Alternatively, when a job seeker enters an input to select a specific major category using a user interface, jobs (competency units) belonging to the major category may be recommended based on the job classification filter for only the selected major category. In the above, the case of recommending jobs by creating various job classification filters for each major category has been described as an example, and it is also possible to create a job classification filter for all competency unit documents belonging to all major categories to recommend jobs. The second job classification filter 240 may not represent a relationship between NCS competency units, but may also represent a relationship between NCS subcategories. In this case, the competency unit documents belonging to the sub-classification are integrated into one document, and a second job classification filter 240 indicating the similarity between sub-classifications may be generated based on the similarity between the integrated documents of sub-classifications.

According to an embodiment of the present invention, a job classification filter is not created by learning using careers (past job experience), such as a job seeker's resume, portfolio, and self-introduction, but based on only job competency standard classification data. By creating a classification filter and recommending jobs to job seekers, it is possible to prevent error factors from occurring in the job classification filter due to uncertain learning data, and it is suitable for job seekers based on the job classification filter prepared with objective, reasonable, and consistent standards. Can recommend a job.

14 is a block diagram of a second job recommendation unit constituting a job recommendation device according to another embodiment of the present invention. 12 and 14, the second job recommendation unit 340 analyzes the job competency criteria related to recruitment based on the recruitment-related data 600 and the second job classification filter 240 to provide one or more job seekers. Recommend employment information. The second job recommendation unit 340 may include a word extracting unit 348 and a job capability criterion analysis unit 350.

The word extractor 348 may extract job-related words related to recruitment from the recruitment-related data 600 of the recruiter. The word extracting unit 348 may include a morpheme analyzer that divides sentences in the recruitment-related data 600 into words in morpheme units and tags parts of speech. In an embodiment, the word extracting unit 348 extracts specific morphemes and morpheme combinations such as nouns and roots after morpheme analysis, processes a self-made stopword list, and applies N-grams such as Bigram in the case of noun continuation can do. The word extractor 348 may extract only words related to the competency unit from among the words of the recruitment-related data 600.

The job competency criteria analysis unit 350 determines one or more third competency units related to the recruiter from the words extracted from the recruitment-related data 600, and determines the third competency unit and the third competency unit based on the second job classification filter 240. One or more fourth capacity units having a high degree of similarity may be extracted. As another example, when the level score required for the third competency unit is input by the recruiter, or when completion information such as career or education programs required in relation to the third competency unit is input, the job competency criteria analysis unit 350 Is to reflect the level score or completion information required for the third competency unit as the level score or completion information required for other competency units based on the similarity matrix relationship between the competency units of the second job classification filter 240. I can. Accordingly, based on the second job classification filter 240, it is possible to expand the hiring-related history requested by the recruiter for the first competency unit to the level required for the other competency units. Job seekers who meet the required level can be recommended to recruiters.

When one or more competency units related to the recruiter are determined, employment information is provided to the job seeker matching the competency unit information of the corresponding employee, so that job matching between the job seeker and the recruiter can be made based on the competency unit. Therefore, recruiters can be recommended for talents that meet the job aptitudes of companies and public institutions, and recruiters are provided with recruitment information that can perform appropriate jobs based on competency units, thereby preventing job mismatch between recruiters and recruiters. Can be prevented.

At least some of the job recommendation methods according to an embodiment of the present invention may be written in a program that can be executed on a computer, and may be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium. Computer-readable recording media include volatile memories such as SRAM (Static RAM), DRAM (Dynamic RAM), SDRAM (Synchronous DRAM), Read Only Memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Nonvolatile memory such as Electrically Erasable and Programmable ROM (EEPROM), flash memory device, phase-change RAM (PRAM), magnetic RAM (MRAM), resistive RAM (RRAM), ferroelectric RAM (FRAM), floppy disk, hard disk, or The optical reading medium may be, for example, a storage medium in the form of a CD-ROM or a DVD, but is not limited thereto.

At least some of the job recommendation methods according to an embodiment of the present invention may be implemented by software (eg, a program) including instructions stored in a computer-readable storage medium. At least some of the job recommendation methods according to an embodiment of the present invention may be implemented by calling a stored command from a storage medium and executing a process according to the called command. Instructions may include code generated or executed by a compiler or interpreter. The storage medium may be provided in the form of a non-transitory storage medium. Here,'non-transient' means that the storage medium does not contain a signal and is tangible, but does not distinguish between semi-permanent or temporary storage of data in the storage medium.

It should be understood that the above embodiments have been presented to aid the understanding of the present invention, and do not limit the scope of the present invention, and various deformable embodiments are also within the scope of the present invention. The technical protection scope of the present invention should be determined by the technical spirit of the claims, and the technical protection scope of the present invention is not limited to the literal description of the claims per se, but the invention of the scope of which the technical value is substantially equal. It should be understood that it extends to.

10: job recommendation device
20: job seeker terminal
30: job recommendation server
40: database
100: job classification filter generation unit
120: first job classification filter generation unit
140: second job classification filter generation unit
160: job classification filter update unit
200: job classification filter
220: first job classification filter
240: second job classification filter
300: job recommendation department
320: 1st job recommendation section
340: second job recommendation unit
400: Job competency standard classification data
500: competency-related data
600: hiring data

Claims (24)

A database for storing a job classification filter generated based on job competency standard classification data including competency unit documents related to competency units; And
It includes a job recommendation unit for recommending a job to the job seeker based on the job seeker's competency-related data and the job classification filter,
The job classification filter includes at least one of a degree of association between the competency unit competency units and words extracted from the competency unit documents and a degree of association between the competency units. The method of claim 1,
The job classification filter includes at least one of a first job classification filter and a second job classification filter,
The first job classification filter includes a degree of association between the competency units and words extracted from the competency unit documents,
The second job classification filter is a job recommendation device indicating a degree of relation between the competency units. The method of claim 2,
The job recommendation unit recommends a job to the job seeker based on the competency-related data and the first job classification filter,
The first job classification filter is a two-dimensional matrix in which competency units are arranged in a first unit of a row unit and a column unit, and words extracted from the competency unit documents are arranged in the other second unit. Is provided,
The two-dimensional matrix is a job recommendation device having, as each component value, a degree of association between a competency unit document associated with each competency unit of the first unit and each word of the second unit. The method of claim 3,
The job recommendation unit,
A word extraction unit for extracting competency related words corresponding to words arranged in the second unit of the first job classification filter from among words of the competency related data;
A vectorization unit for generating a feature vector by vectorizing the competency-related words according to an arrangement order of words arranged in the second unit of the first job classification filter;
A similarity calculation unit for calculating a similarity between the feature vector and a competency unit vector corresponding to each competency unit of the first job classification filter; And
A job recommendation device comprising a recommended job determination unit configured to determine a recommended job in order of a high degree of similarity between competency unit vectors corresponding to the competency units and the feature vector. The method of claim 3,
The job recommendation unit classifies jobs related to recruitment based on recruitment-related data and the first job classification filter, and recommends one or more employment information to the job seeker based on the job classification information of the recruiter,
The job recommendation unit,
A word extracting unit for extracting recruitment-related words corresponding to words arranged in the second unit of the first job classification filter from among words of the recruitment-related data;
A vectorization unit that vectorizes the recruitment-related words according to an arrangement order of words arranged in the second unit of the first job classification filter to generate a feature vector;
A similarity calculation unit for calculating a similarity between the feature vector and a competency unit vector corresponding to each competency unit of the first job classification filter; And
A job recommendation apparatus comprising a job classification unit configured to determine one or more competency units related to the recruiter in the order of high similarity between competency unit vectors corresponding to the competency units and the feature vector. The method of claim 2,
The job recommendation unit recommends a job to the job seeker based on the competency-related data and the second job classification filter,
The second job classification filter includes a two-dimensional matrix in which competency units are arranged in a first unit of a row unit and a column unit, and the competency units are arranged in another second unit,
The two-dimensional matrix is a job recommendation device having a relationship between a first competency unit document associated with each competency unit of a first unit and a second competency unit document associated with each competency unit of the second unit as respective component values. The method of claim 6,
The job recommendation unit,
A word extraction unit for extracting words related to the competency unit of the job seeker from the competency related data;
A job competency analysis unit that determines one or more first competency units related to the job seeker from words extracted from the competency-related data; And
A job recommendation device including a recommended job determination unit configured to determine at least one second competency unit among a plurality of competency units as a recommended job based on a similarity with the first competency unit based on the second job classification filter. The method of claim 7,
The recommended job determination unit applies the job-seeker's competency-related history information related to the first competency unit to the second job classification filter and reflects the competency-related history information for other competency units. The method of claim 6,
The job recommendation unit recommends one or more recruitment information to the job seeker by analyzing the job competency criteria related to recruitment based on the recruitment data and the second job classification filter,
The job recommendation unit,
A word extracting unit for extracting words associated with a competency unit from the hiring-related data; And
One or more first competency units related to the recruiter are determined based on words extracted from the recruitment-related data, and among a plurality of competency units in order of similarity to the first competency unit based on the second job classification filter A job recommendation device including a job competency criterion analysis unit that determines one or more second competency units. The method of claim 2,
Further comprising a job classification filter generating unit that generates at least one of the first job classification filter and the second job classification filter based on the job competency standard classification data,
The job classification filter generation unit includes at least one of a first job classification filter generation unit that generates the first job classification filter and a second job classification filter generation unit that generates the second job classification filter,
The first job classification filter generating unit generates the first job classification filter by calculating a degree of association between the competency unit documents and words extracted from the competency unit documents,
The second job classification filter generation unit generates the second job classification filter by calculating a relationship between the competency unit documents based on words extracted from the competency unit documents. The method of claim 10,
The job recommendation unit recommends a job to the job seeker based on the competency-related data and the first job classification filter generated by the first job classification filter generation unit,
The first job classification filter generation unit,
A document extracting unit that extracts NCS competency unit documents for each major category among the classification data of National Competency Standards (NCS);
A word extracting unit for extracting words from all of the NCS capability unit documents belonging to the same major classification;
Association between NCS capability units and the extracted words based on the word frequency inverse document frequency (TF-IDF) between words extracted from the NCS capability unit documents and the NCS capability unit documents A correlation analysis unit for analyzing a degree; And
A job recommendation apparatus comprising a normalization unit for normalizing a size of a 2D matrix having a degree of association between the NCS capability units analyzed by the association analysis unit and the extracted words as component values to be a set value. The method of claim 11,
The document extracting unit is a job recommendation device for extracting a document including names, definitions, basic vocational skills, evaluation items, performance criteria for each competency unit element belonging to the NCS competency unit, knowledge, skills, and attitude items for each NCS competency unit. The method of claim 10,
The job recommendation unit recommends a job to the job seeker based on the competency-related data and the second job classification filter generated by the second job classification filter generation unit,
The second job classification filter generation unit,
A document extracting unit for extracting NCS competency unit documents from the classification data of National Competency Standards (NCS);
A word extracting unit that extracts words from all of the NCS capability unit documents; And
A job recommendation device comprising a classification relationship analysis unit that analyzes a degree of association between NCS competency unit documents based on the similarity between words extracted from the NCS competency unit documents. The method of claim 1,
The job recommendation unit,
A job recommendation device for displaying the recommended job in a different display manner depending on whether the recommended job determined by the job recommendation unit matches the job selected by the job seeker. The method of claim 1,
The competency-related data includes education information, certificate information, portfolio, resume, self-introduction, SNS data, and data collected by chatbots that collect history information in the form of conversations with the job seeker, user's social relationship information, SNS interests, Job recommendation device including digitized information of job seekers such as content viewing records and search information. Recommending a job to the job seeker based on the job-seeker's competency-related data and a job classification filter by a job recommendation unit,
The job classification filter is a filter generated based on competency unit documents related to competency units of job competency standard classification data,
The job classification filter includes at least one of a degree of association between the competency unit and competency units and words extracted from the competency unit documents, and a degree of association between the competency units. The method of claim 16,
The job classification filter includes a two-dimensional matrix in which competency units are arranged in a first unit of a row unit and a column unit, and words extracted from the competency unit documents are arranged in the other second unit, and ,
The two-dimensional matrix is a job recommendation method having a degree of association between a competency unit document associated with each competency unit of the first unit and each word of the second unit as respective component values. The method of claim 17,
The step of recommending the job,
Extracting, by a word extracting unit, competency related words corresponding to words arranged in the second unit of the job classification filter from among words of the competency related data;
Generating a feature vector by vectorizing the competency-related words according to an arrangement order of words arranged in the second unit of the job classification filter by a vectorization unit;
Calculating a similarity between the feature vector and the capability unit vector corresponding to each capability unit of the job classification filter, by a similarity calculation unit; And
A job comprising the step of determining, by a recommended job determination unit, a recommended job including at least one competency unit among the competency units in the order of high similarity between competency unit vectors corresponding to the competency units and the feature vector Recommended way. The method of claim 16,
The job classification filter includes a two-dimensional matrix in which competency units are arranged in a first unit of a row unit and a column unit, and the competency units are arranged in the other second unit,
The two-dimensional matrix is a method of recommending a job having a degree of relationship between a first competency unit document associated with each competency unit of the first unit and a second competency unit document associated with each competency unit of the second unit as respective component values . The method of claim 19,
The step of recommending the job,
Extracting, by a word extracting unit, words related to the competency unit of the job seeker from the competency related data;
Determining one or more first competency units related to the job seeker from words extracted from the competency-related data, by a job competency analysis unit; And
A job recommendation method comprising the step of determining, by the job classification filter, at least one second competency unit from among a plurality of competency units as a recommended job based on the similarity with the first competency unit, by the job classification filter . The method of claim 16,
Further comprising the step of generating the job classification filter based on the job competency standard classification data,
The step of generating the job classification filter,
Extracting NCS competency unit documents for each major classification among the classification data of National Competency Standards (NCS) by a document extraction unit;
Extracting, by a word extracting unit, words from all of the NCS capability unit documents belonging to the same major classification;
By the association analysis unit, based on the word frequency inverse document frequency (TF-IDF; Term Frequency Inverse Document Frequency) between the words extracted from the NCS capability unit documents and the NCS capability unit documents, the NCS capability units and Analyzing a degree of association between the extracted words; And
A job recommendation method comprising the step of dividing, by a normalization unit, a 2D matrix having a degree of association between the NCS capability units analyzed by the association analysis unit and the extracted words as component values, and normalizing them by a set value. . The method of claim 16,
Further comprising the step of generating the job classification filter based on the job competency standard classification data,
The step of generating the job classification filter,
Extracting, by the document extracting unit, NCS competency unit documents from National Competency Standards (NCS) classification data;
Extracting, by a word extracting unit, words from all of the NCS capability unit documents; And
A job recommendation method comprising the step of analyzing, by a relationship between classifications analysis unit, a degree of association between NCS competency unit documents based on the similarity of words extracted from the NCS competency unit documents. The method of claim 16,
The job classification filter updating unit further comprises the step of collecting update information of the job competency standard classification data by a crawling technique, and updating the job classification filter based on the update information of the job competency standard classification data Recommended way. A computer-readable recording medium in which a program for executing the job recommendation method according to any one of claims 16 to 23 is recorded.

Images