A candidate determination method and device based on knowledge graph

Abstract

The present application provides a candidate determination method and device based on a knowledge graph, wherein the candidate determination method based on the knowledge graph includes: constructing a position skill knowledge graph and a candidate skill knowledge graph; obtaining recruitment demand information, and generating a recruitment skill subgraph based on the recruitment demand information; determining the candidate positions corresponding to the recruitment demand information according to the recruitment skill subgraph and the position skill knowledge graph; determining the candidate set corresponding to the recruitment demand information according to the candidate positions and the candidate skill knowledge graph.

Description

Technical Field

[0001] The present application relates to the field of knowledge graph technology, and in particular to a candidate determination method and apparatus based on a knowledge graph, a computing device, and a computer-readable storage medium. Background Art

[0002] As companies grow and develop, the demand for talent is getting higher and higher. Currently, corporate recruiters usually send out recruitment information and discover or screen candidates based on the needs of the employing department to complete the recruitment.

[0003] Existing talent recruitment methods generally rely on manual experience, analyze employment needs, define positions, and screen candidates; or use intelligent methods such as machine learning to extract features from resumes and recruitment information for similarity matching to determine candidates. However, methods based on manual experience are highly subjective, and different recruiters will have different results. They are inefficient and prone to errors. Intelligent methods based on machine learning can only match the similarity between resumes and recruitment information, lack matching between different positions and different candidates, and cannot recommend similar positions or similar candidates. The recommendation efficiency is low and the effect is not ideal. Summary of the invention

[0004] In view of this, the embodiments of the present application provide a candidate determination method and apparatus based on a knowledge graph, a computing device, and a computer-readable storage medium to address the technical defects existing in the prior art.

[0005] According to a first aspect of an embodiment of the present application, a candidate determination method based on a knowledge graph is provided, comprising:

[0006] Build job skill knowledge graphs and candidate skill knowledge graphs;

[0007] Acquire recruitment demand information, and generate a recruitment skill subgraph based on the recruitment demand information;

[0008] Determine the candidate positions corresponding to the recruitment demand information according to the recruitment skill subgraph and the position skill knowledge graph;

[0009] According to the candidate positions and the candidate skill knowledge graph, a candidate set corresponding to the recruitment demand information is determined.

[0010] According to a second aspect of an embodiment of the present application, a candidate determination device based on a knowledge graph is provided, comprising:

[0011] A building module, configured to build a job skill knowledge graph and a candidate skill knowledge graph;

[0012] A generating module is configured to obtain recruitment demand information and generate a recruitment skill subgraph based on the recruitment demand information;

[0013] A position determination module is configured to determine a candidate position corresponding to the recruitment demand information according to the recruitment skill subgraph and the position skill knowledge graph;

[0014] The candidate determination module is configured to determine a candidate set corresponding to the recruitment demand information based on the candidate positions and the candidate skill knowledge graph.

[0015] According to a third aspect of an embodiment of the present application, a computing device is provided, comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein the processor implements the steps of the candidate determination method based on the knowledge graph when executing the computer instructions.

[0016] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores computer instructions, and when the computer instructions are executed by a processor, the steps of the candidate determination method based on the knowledge graph are implemented.

[0017] According to a fifth aspect of an embodiment of the present application, a chip is provided, which stores computer instructions, and when the computer instructions are executed by the chip, the steps of the candidate determination method based on the knowledge graph are implemented.

[0018] In the embodiment of the present application, by constructing a job skill knowledge graph and a candidate skill knowledge graph and generating a recruitment skill subgraph based on recruitment demand information, the candidate job corresponding to the recruitment demand information is determined according to the recruitment skill subgraph and the job skill knowledge graph, and the candidate set corresponding to the recruitment demand information is determined according to the candidate job and the candidate skill knowledge graph. The candidate determination method of the present application first determines the candidate job based on the knowledge graph, and then determines the candidate based on the candidate job. The recommendation efficiency is high, and the recommended candidate has a high matching degree with the job, which is convenient for the employer to efficiently select candidates. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a structural block diagram of a computing device provided in an embodiment of the present application;

[0020] Figure 2 is a flow chart of a candidate determination method based on a knowledge graph provided in an embodiment of the present application;

[0021] Figure 3 It is a schematic diagram of the job skill knowledge graph provided in the embodiment of the present application;

[0022] Figure 4 It is a schematic diagram of the candidate skill knowledge graph provided in the embodiment of the present application;

[0023] Figure 5 It is a schematic diagram of a recruitment skill subgraph provided in an embodiment of the present application;

[0024] Figure 6 This is a flow chart of a candidate determination method provided by an embodiment of the present application, taking Company A's recruitment for an operations position as an example;

[0025] Figure 7 This is a schematic diagram of a job skill knowledge graph provided by another embodiment of the present application;

[0026] Figure 8 is a schematic diagram of a candidate skill knowledge graph provided by another embodiment of the present application;

[0027] Fig. 9 is a schematic diagram of a recruitment skill subgraph provided by another embodiment of the present application;

[0028] Fig.10 It is a structural diagram of a candidate determination device based on a knowledge graph provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0030] The terms used in one or more embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of the present application. The singular forms of "a", "said" and "the" used in one or more embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of the present application refers to and includes any or all possible combinations of one or more associated listed items.

[0031] It should be understood that, although the terms first, second, etc. may be used to describe various information in one or more embodiments of the present application, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "in response to determination".

[0032] First, the terms involved in one or more embodiments of the present invention are explained.

[0033] Knowledge Graph: It is a graph-based data structure consisting of nodes and edges. Each node represents an "entity" and each edge represents a "relationship" between entities. The knowledge graph is essentially a semantic network. Entities can refer to things in the real world, such as people, places, companies, telephones, animals, etc.; relationships are used to express certain connections between different entities.

[0034] Subgraph: Subgraph is one of the basic concepts in graph theory, which refers to a graph whose node set and edge set are subsets of the node set and edge set of a certain graph respectively.

[0035] Clustering algorithm: Cluster analysis, also known as group analysis, is a statistical analysis method for studying (sample or indicator) classification problems, and is also an important algorithm for data mining. Cluster analysis is composed of several patterns. Usually, a pattern is a vector of measurement or a point in multidimensional space. Cluster analysis is based on similarity. Patterns in a cluster have more similarities than patterns in different clusters.

[0036] Community: A community is a subgraph that contains nodes and edges. The nodes in the same community are closely related, while the connections between communities are relatively sparse.

[0037] Community Detection Algorithm: A computational method for clustering tightly structured nodes in a network.

[0038] In the present application, a candidate determination method and apparatus based on a knowledge graph, a computing device, and a computer-readable storage medium are provided, which are described in detail one by one in the following embodiments.

[0039] Figure 1 The structure block diagram of a computing device 100 according to an embodiment of the present application is shown. The components of the computing device 100 include but are not limited to a memory 110 and a processor 120. The processor 120 is connected to the memory 110 via a bus 130, and the database 150 is used to store data.

[0040] The computing device 100 also includes an access device 140 that enables the computing device 100 to communicate via one or more networks 160. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 140 may include one or more of any type of network interface (e.g., a network interface card (NIC)) whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a World Wide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.

[0041] In one embodiment of the present application, the above components of the computing device 100 and Figure 1 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 1 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of the present application. Those skilled in the art may add or replace other components as needed.

[0042] The computing device 100 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. The computing device 100 may also be a mobile or stationary server.

[0043] The processor 120 may execute Figure 2 The steps in the candidate determination method based on the knowledge graph are shown. Figure 2 A flowchart of a candidate determination method based on a knowledge graph according to an embodiment of the present application is shown, including steps 202 to 208.

[0044] Step 202: Construct a job skill knowledge graph and a candidate skill knowledge graph.

[0045] The job skill knowledge graph can be constructed based on the employer's requirements for the job. Since different employers may have different skill requirements for the same job, the job skill knowledge graph can be constructed based on the employer's actual needs.

[0046] The candidate skill knowledge graph can be constructed based on the resume information in the employer's talent pool. After the recruitment process is initiated, for the received resumes, the candidate information and the skill information of the candidate corresponding to each resume are identified, and the candidate skill knowledge graph is constructed based on the candidate information and the corresponding skill information.

[0047] In practical applications, the specific steps for building a job skills knowledge graph are:

[0048] Obtain job skills data;

[0049] Parsing the job skill data to obtain job information, skill information corresponding to the job, and job skill weight information;

[0050] A job skill knowledge graph is constructed using the job information and the skill information corresponding to the job as nodes and the job skill weight information as a relationship.

[0051] Specifically, the employer's demand for positions can be summarized in a table to obtain a position skill data table containing multiple positions skill data. Identify the data in the position skill data table and obtain the information required to build a knowledge graph, such as: position information, skill information corresponding to the position, and weight information of the position skills. The obtained position information and skill information corresponding to the position are used as node data for building a position skill knowledge graph, and the obtained position skill weight information is used as relationship data for building a position skill knowledge graph. The constructed position skill knowledge graph contains position nodes, skill nodes corresponding to the position, and skill weight relationships.

[0052] In a specific implementation of this application, the positions of secretary, administrative assistant and algorithm engineer are used as examples to describe the construction of a job skill knowledge graph. In this embodiment, a directed weighted graph is used, with the positions of the employer and the skills corresponding to the positions as nodes, and the skill weights as relationships. The job skill knowledge graph is constructed as follows: Figure 3 The diagram of the job skill knowledge graph is shown as follows. For secretary positions, it is necessary to master the skills of office software, official document writing and event organization, among which the skill weight of office software is 0.5, the skill weight of official document writing is 0.4, and the skill weight of event organization is 0.1; for administrative assistant positions, it is necessary to master the skills of office software, event organization and work summary, among which the skill weight of office software is 0.4, the skill weight of event organization is 0.3, and the skill weight of work summary is 0.3; for algorithm engineer positions, it is necessary to master the skills of business understanding, data cleaning and data mining, among which the skill weight of business understanding is 0.4, the skill weight of data cleaning is 0.4, and the skill weight of data mining is 0.3.

[0053] In practical applications, the specific steps to build a candidate skill knowledge graph are:

[0054] Obtain candidate resumes from a candidate resume database;

[0055] Parsing the candidate resume to obtain candidate information, candidate corresponding skill information and candidate skill weight information;

[0056] The candidate information and the skill information corresponding to the candidate are used as nodes, and the candidate skill weight information is used as a relationship to construct a candidate skill knowledge graph.

[0057] Specifically, an employer will receive multiple resumes of applicants during each recruitment process, and the resumes can be added to the employer's candidate resume database. When a suitable candidate needs to be found, the candidate resume can be obtained from the candidate resume database. It should be noted that the candidate resume database can store the candidate's resume file or the candidate's resume information, which is the candidate information extracted after identifying the candidate's resume file.

[0058] When the candidate resume database stores the candidate's resume file, the resume information in the candidate resume file is identified and extracted to obtain the candidate information in the resume, the candidate's corresponding skill information, and the candidate's skill weight information; when the candidate resume database stores the candidate's resume information, the candidate information of the candidate, the candidate's corresponding skill information, and the candidate's skill weight information can be directly obtained.

[0059] The candidate information and the skill information corresponding to the candidate are used as the node information for constructing the candidate skill knowledge graph, and the candidate skill weight information is used as the relationship information for constructing the candidate skill knowledge graph. The constructed candidate skill knowledge graph contains the candidate node, the skill node corresponding to the candidate, and the candidate skill weight relationship.

[0060] In a specific implementation of the present application, taking the candidates in the talent pool as Zhang San, Li Si and Wang Wu as an example, the construction of the candidate skill knowledge graph is described in detail. Figure 4 As shown in the schematic diagram of the candidate skill knowledge graph, a directed weighted graph is used with candidates and their skills as nodes and the weights of the skills as relationships. The candidate skill knowledge graph is constructed as follows: Figure 4As shown. For candidate Zhang San, he has the skills of office software, official document writing and event organization, and the skill weight of office software is 0.6, the skill weight of official document writing is 0.3, and the skill weight of event organization is 0.1; for Li Si, he has the skills of office software, event organization and work summary, among which the skill weight of office software is 0.4, the skill weight of event organization is 0.2, and the skill weight of work summary is 0.4. For candidate Wang Wu, he has the skills of business understanding, data cleaning and data mining, among which the skill weight of business understanding is 0.4, the skill weight of data cleaning is 0.4, and the skill weight of data mining is 0.3.

[0061] Build a job skill knowledge graph based on the employer's skill requirements for the position, and build a candidate skill knowledge graph based on the resumes in the talent pool, so as to quickly find candidate resumes that meet the recruitment needs in the talent pool.

[0062] Step 204: Obtain recruitment demand information, and generate a recruitment skill subgraph based on the recruitment demand information.

[0063] The recruitment demand information includes the target position name, the skills corresponding to the target position, and the skill weight of the target position. The employer may recruit multiple positions at the same time. The recruitment demand information is the recruitment demand of one of the multiple positions, that is, each recruitment position can generate a recruitment skill subgraph according to its corresponding recruitment demand. For example, the company is recruiting two positions, a secretary position and a sales position. According to the recruitment demand corresponding to the secretary position, a secretary recruitment skill subgraph is generated, and according to the recruitment demand corresponding to the sales position, a sales recruitment skill subgraph is generated, and two recruitment skill subgraphs corresponding to the two positions are obtained.

[0064] In practical applications, methods for constructing a recruitment skill subgraph based on recruitment demand information include:

[0065] Analyze the recruitment demand information to obtain target position information, skill information corresponding to the target position, and skill weight information of the target position;

[0066] A recruitment skill subgraph is generated with the target position information and the skill information corresponding to the target position as nodes and the target position skill weight information as a relationship.

[0067] Specifically, one recruitment requirement is selected from multiple recruitment requirements as recruitment requirement information, and the recruitment requirement information includes a recruitment position, skills corresponding to the recruitment position, and skill weights. The recruitment requirement information is parsed to obtain target position information, skill information corresponding to the target position, and target position skill weight information in the recruitment requirement information. The target position information and the skill information corresponding to the target position are used as node data for generating a recruitment skill subgraph, and the target position skill weight information is used as relationship data for generating a recruitment skill subgraph, thereby generating a recruitment skill subgraph.

[0068] In a specific implementation manner of the present application, Figure 5 Taking the recruitment skill subgraph diagram as an example, the generation of the recruitment skill subgraph is described in detail. In this embodiment, a directed weighted graph is used with recruitment positions and corresponding skills as nodes and skill weights as relationships to construct the following structure: Figure 5 The recruitment skills subgraph shown in the schematic diagram of the recruitment skills subgraph. According to the recruitment needs, it can be determined that the positions to be recruited require skills in office software, event organization, official document writing and work summary, among which the weight of office software skills is 0.5, the weight of event organization skills is 0.3, the weight of official document writing skills is 0.1, and the weight of work summary skills is 0.1.

[0069] By constructing a recruitment skill sub-graph according to recruitment needs, it is convenient to select candidate resumes that meet the requirements from the employer's talent pool based on the recruitment skill sub-graph.

[0070] Step 206: Determine the candidate positions corresponding to the recruitment demand information according to the recruitment skill subgraph and the position skill knowledge graph.

[0071] The recruitment skill subgraph includes the target job title and the skills corresponding to the target job title, and the job skill knowledge graph includes the job title and the skills corresponding to the job title. The skills corresponding to the target job title and the skills corresponding to the job title are compared to determine one or more jobs in the job skill knowledge graph that match the job title in the recruitment subgraph.

[0072] In practical applications, the candidate positions corresponding to the recruitment demand information are determined based on the graph similarity between the recruitment skill subgraph and the position skill knowledge graph.

[0073] The graph similarity between the recruitment skill subgraph and the position skill knowledge graph refers to the similarity between the skill node in the recruitment skill subgraph and the skill node in the position skill knowledge graph.

[0074] The steps for determining the candidate positions corresponding to the recruitment demand information based on the graph similarity between the recruitment skill subgraph and the job skill knowledge graph are as follows:

[0075] Converting the recruitment skill subgraph into a recruitment skill vector, and converting the job skill knowledge graph into multiple job skill vectors;

[0076] Calculating the job similarity between the recruitment skill vector and each job skill vector respectively;

[0077] Determine a candidate position corresponding to the recruitment requirement information according to the position similarity.

[0078] Specifically, the recruitment skill subgraph and the position skill knowledge graph are converted into recruitment skill vectors and position skill vectors, that is, the recruitment skill subgraph and the position skill knowledge graph are converted into vectorized form. The position skill knowledge graph contains multiple position information. After each position information is vectorized, multiple position skill vectors can be obtained. The similarity between the recruitment skill vector and each position skill vector is calculated respectively, so as to determine the position corresponding to the recruitment demand information in the position skill knowledge graph. Specific methods for calculating similarity include but are not limited to Euclidean distance, Manhattan distance, Minkowski distance, cosine similarity, and Pearson Correlation Coefficient.

[0079] In practical applications, the candidate positions corresponding to the recruitment demand information are determined according to the position similarity, including but not limited to the following two methods:

[0080] Sort the positions according to the position similarities, and select a preset number of positions to generate candidate positions as candidate positions corresponding to the recruitment demand information; or

[0081] The positions whose similarity to the positions is greater than a preset threshold are selected as candidate positions corresponding to the recruitment demand information.

[0082] In a specific embodiment of the present application, the steps of determining the candidate positions corresponding to the recruitment demand information are further explained by taking the example of sorting the candidate positions according to similarity. After calculating the similarity between each position in the position skill knowledge graph and the position in the recruitment skill subgraph, sort each position in the position skill knowledge graph according to the similarity value. According to the actual needs of the recruitment, multiple positions with the highest similarity are selected as candidate positions corresponding to the recruitment needs.

[0083] In another specific embodiment of the present application, taking a preset similarity threshold as an example, the step of determining the candidate positions corresponding to the recruitment demand information is further explained. The similarity threshold is set as the preset threshold according to the actual recruitment demand. After calculating the similarity between each position in the position skill knowledge graph and the position in the recruitment skill subgraph, the positions in the skill knowledge graph with a similarity greater than the preset threshold are taken as candidate positions corresponding to the recruitment demand information.

[0084] Step 208: Determine a candidate set corresponding to the recruitment demand information based on the candidate positions and the candidate skill knowledge graph.

[0085] According to the skills corresponding to the candidate positions, candidates who master the skills are determined in the candidate skill knowledge graph to form a candidate set.

[0086] In practical applications, the steps of determining the candidate set corresponding to the recruitment demand information according to the candidate positions and the candidate skill knowledge graph are as follows:

[0087] Determining a skill information subgraph corresponding to the candidate position in the candidate skill knowledge graph based on the candidate position;

[0088] According to the skill information subgraph corresponding to the candidate position, a candidate set corresponding to the recruitment demand information is determined in the candidate skill knowledge graph.

[0089] Specifically, the candidate position corresponds to a position skill, and the position skill corresponding to the candidate position is obtained in the candidate skill knowledge graph to construct a skill information subgraph, wherein the skill information subgraph includes a candidate node, a skill node corresponding to the candidate, and a skill weight relationship. According to the skill node corresponding to the candidate position, the candidate node connected to the skill node is determined, so as to determine the candidate corresponding to the candidate position in the candidate knowledge graph, and the candidates form a candidate set.

[0090] In practical applications, according to the skill information subgraph corresponding to the candidate position, the specific steps of determining the candidate set corresponding to the recruitment demand information in the candidate skill knowledge graph are as follows:

[0091] Converting the skill information subgraph corresponding to the candidate position into a candidate position skill vector, and converting the candidate skill knowledge graph into a candidate skill vector;

[0092] Calculating the skill similarity between the candidate position skill vector and the candidate skill vector respectively;

[0093] A candidate set corresponding to the recruitment requirement information is determined according to the skill similarity.

[0094] Specifically, the skill information subgraph and the candidate skill knowledge graph are converted into candidate position skill vectors and candidate skill vectors, that is, the skill information subgraph and the candidate skill knowledge graph are converted into vector form. The candidate skill knowledge graph contains multiple candidate information, and each candidate information is vectorized to obtain multiple candidate skill vectors. The skill similarity between the candidate position skill vector and each candidate skill vector is calculated respectively, so as to determine the candidate set corresponding to the recruitment demand information in the candidate skill knowledge graph. Specific methods for calculating similarity include but are not limited to Euclidean distance, Manhattan distance, Minkowski distance, cosine similarity, and Pearson Correlation Coefficient.

[0095] In practical applications, a set of candidates corresponding to the recruitment demand information is determined based on the skill similarity, including but not limited to the following two methods:

[0096] Sort the candidates according to the skill similarity, and select a preset number of candidates to generate a candidate set; or

[0097] Positions whose skill similarity is greater than a preset threshold are selected to generate a candidate set.

[0098] Specifically, methods for determining a candidate set based on skill similarity include but are not limited to the above two methods, and the skill similarity can be used according to actual conditions.

[0099] In a specific embodiment of the present application, a method for determining a candidate set based on skill similarity is further explained by taking the sorting of candidates based on skill similarity as an example. The skill similarity between the skill corresponding to the candidate position and the skill corresponding to each candidate is calculated, and the candidates are sorted in descending order based on the skill similarity. A preset number of candidates are selected based on the actual recruitment situation, and the preset number of candidates constitute a candidate set.

[0100] In another specific embodiment of the present application, a method of determining a candidate set based on skill similarity is further explained by taking a method of presetting a similarity threshold as an example. A similarity threshold is pre-set, and a specific similarity threshold can be set according to the recruitment needs of the employer, such as a similarity of 90%. The skill similarity between the skills corresponding to the candidate position and the skills corresponding to the candidate is calculated, and candidates whose skill similarity is greater than the preset threshold are selected to generate a candidate set.

[0101] In practical applications, the candidate set can also be determined through the clustering algorithm of the knowledge graph. The specific steps are as follows:

[0102] Cluster candidates according to their skills to generate multiple skill communities;

[0103] Determining a target skill community among the multiple skill communities according to the skill information subgraph;

[0104] The candidates in the target skill community are obtained to generate a candidate set.

[0105] Specifically, the candidate skill knowledge graph is clustered in advance according to the community detection algorithm, and the community results are obtained according to the skills mastered by each candidate. The community detection algorithm is used to discover the community structure in the network. The community structure is a common feature of the knowledge graph network. The entire network is composed of many communities.

[0106] In the community discovery algorithm, it is impossible to determine the number of communities in advance. A measurement method is needed to measure whether each result is the best result in the calculation process. This measurement method is modularity. Modularity is used to measure whether the division of a community is a relatively good result. A relatively good result has a high similarity of nodes within the community and a low similarity of nodes outside the community. The size of modularity is defined as the ratio of the total number of edges within the community to the total number of edges in the network minus an expected value. The expected value is the ratio of the total number of edges within the community to the total number of edges in the network formed by the same community allocation when the network is set as a random network. The algorithm of modularity is shown in the following formula 1:

[0107]

[0108] Among them, Q represents modularity. The larger the modularity, the better the community division effect. The range of Q is [-0.5, 1). When the Q value is between 0.3 and 0.7, it means that the clustering effect in the knowledge graph is good. Suppose there are x nodes in the knowledge graph, each node represents an input, and we divide these inputs into N communities, and there are m connections between the nodes. v and w are any two nodes in x. Avw Indicates whether the two nodes v and w are directly connected. If v and w are directly connected, then A vw =1, if v and w are not directly connected, then A vw =0. k v represents the degree of node v, k w represents the degree of node w. Degree is the basic knowledge of graph theory. If there are several edges from a node, we say the degree of this node is how much. v , c w ) is used to determine whether v and w are in the same community. If so, then δ(c v , c w )=1, if not, then δ(c v , c w )=0.

[0109] After clustering the candidate skill knowledge graph, multiple communities will be obtained. Each community includes candidates with the same or similar skills. Then, the candidates are matched in the candidate skill knowledge graph according to the skill information corresponding to the candidate positions. The corresponding target community is matched for the candidate positions according to the skill information, and the candidates corresponding to the target community are taken as the candidate set.

[0110] In a specific embodiment of the present application, after the clustering algorithm of the above-mentioned knowledge graph determines the candidate set, the similarity between the skills in the recruitment skill subgraph and the skills of each candidate in the candidate set can be calculated respectively, and the candidates with higher skill similarity can be further determined based on the similarity to generate a candidate set that meets the similarity requirements. The specific method of determining the candidate set based on the similarity can refer to the above-mentioned preset similarity threshold or the method of sorting according to the similarity.

[0111] By constructing a knowledge graph, we can generate a job skill knowledge graph and a candidate skill knowledge graph, normalize skills, and normalize different identifiers of the same skill to effectively reflect the relationship between positions, skills, and candidates, which is conducive to obtaining similar positions and similar candidates, solving the problems of low efficiency and strong subjectivity of manual screening, and improving work efficiency. Through the method of similarity calculation, we can better screen similar positions and candidates, determine more qualified candidates from different dimensions for the recruitment demand department to select, and effectively improve work efficiency.

[0112] Figure 6 A candidate determination method based on a knowledge graph according to an embodiment of the present application is shown. The method is described by taking Company A's recruitment for an operations position as an example, and includes steps 602 to 608.

[0113] Step 602: Construct a job skill knowledge graph and a candidate skill knowledge graph.

[0114] Specifically, a job skill knowledge graph is constructed based on the employer's requirements for job skills, and the information of resumes in the employer's talent pool is obtained to construct a candidate skill knowledge graph.

[0115] In this embodiment, the employer's requirements for job skills are a job skill table generated by recording and summarizing the requirements of Company A for job skills. The job title information, skill information corresponding to the job, and skill weight information in the table are extracted. A directed weighted graph is used with the job title information and skill information corresponding to the job as node data and skill weight information as relationship data to construct a job skill knowledge graph that meets the job requirements of Company A. The constructed knowledge graph is shown in the figure below. Figure 7 A schematic diagram of a job skill knowledge graph is provided in another embodiment of the present application.

[0116] Figure 7 It includes the positions of new media operation, community operation and event operation. The skills required for new media operation include office software, data analysis and copywriting, with skill weights of 0.3, 0.5 and 0.2 respectively; the skills required for community operation include data analysis, copywriting and communication, with skill weights of 0.4, 0.3 and 0.3 respectively; the skills required for event operation include communication, execution and teamwork, with skill weights of 0.5, 0.3 and 0.2 respectively.

[0117] During the recruitment process, Company A stores the resumes submitted to the company in the talent pool. Obtain the resumes in the talent pool and extract the resume information, which includes the candidate's name information, candidate's skill information, and candidate's skill weight information. Use a directed weighted graph with the candidate's name information and candidate's skill information as node data and the candidate's skill weight information as relationship data to construct a candidate's skill knowledge graph. The constructed knowledge graph is shown in the figure below. Figure 8 A schematic diagram of a candidate skill knowledge graph is provided in another embodiment of the present application.

[0118] Figure 8 The candidates include Liu Yi, Liu Er and Liu San. Liu Yi's skills include office software, data analysis and copywriting, with the weights of 0.3, 0.5 and 0.2 respectively; Liu Er's skills include data analysis, copywriting and communication, with the weights of 0.4, 0.3 and 0.3 respectively; Liu San's skills include communication, execution and teamwork, with the weights of 0.5, 0.3 and 0.2 respectively.

[0119] Step 604: Obtain recruitment demand information, and generate a recruitment skill subgraph based on the recruitment demand information.

[0120] In this embodiment, the recruitment demand information is the recruitment information of the recruitment operation position released by Company A, and the recruitment skill subgraph is constructed based on the recruitment information of the recruitment operation position. The recruitment information of the recruitment operation position contains the position title information, the skill information corresponding to the position, and the skill weight information. With the position title information and the skill information corresponding to the position as the node data and the skill weight information as the relationship data, the following is generated: Fig. 9 The Hiring Skills sub-graph is shown.

[0121] Fig. 9 The skills required for operations positions include copywriting, communication, data analysis and office software, with the skill weights being 0.5, 0.1, 0.2 and 0.2 respectively.

[0122] Step 606: Calculate the graph similarity between the recruitment skill subgraph and the position skill knowledge graph, and determine the candidate positions corresponding to the recruitment demand information based on the graph similarity.

[0123] In this embodiment, it can be known from step 604 that the recruitment demand information is the recruitment information for the operation position released by Company A. Convert the recruitment skill subgraph into a recruitment skill vector, that is, an operation vector; convert the position skill knowledge graph into a position skill vector, that is, a new media operation vector, a community operation vector, and an activity operation vector; calculate the distance between the operation vector and the new media operation vector, the community operation vector, and the activity operation vector respectively, and the value of the distance is the graph similarity between the recruitment skill subgraph and the position skill knowledge graph. Sort the new media operation, community operation, and activity operation positions in the position skill knowledge graph according to the graph similarity, and select a preset number of positions as positions similar to the operation. In this embodiment, the preset number is 2, and new media operation and community operation with high similarity to the operation position are selected as candidate positions according to the preset number. In actual applications, a similarity threshold can also be preset, and positions with a similarity greater than the preset threshold are selected as candidate positions.

[0124] Step 608: Determine a candidate set corresponding to the recruitment requirement information based on the candidate positions and the candidate skill knowledge graph.

[0125] Specifically, a skill information subgraph corresponding to the candidate position is determined in the candidate skill knowledge graph according to the candidate position, the graph similarity between the candidate skill knowledge graph and the skill information subgraph is calculated, and a candidate set corresponding to the recruitment demand information is determined according to the graph similarity.

[0126] In this embodiment, it can be seen from step 608 that the candidate positions of new media operation and community operation are candidate positions. According to the candidate position, the skills matching the candidate position are determined in the candidate skill knowledge graph. Taking the determination of the position skills of new media operation as an example, the skill nodes of copywriting, data analysis and office software are determined in the candidate skill knowledge graph to generate the skill information subgraph of the new media operation position.

[0127] The skill information subgraph is converted into a candidate position skill vector, and the candidate skill knowledge graph is converted into a candidate skill vector. The skill similarity between the candidate position skill vector and the candidate skill vector is calculated, and the candidates Liu Yi and Liu Er corresponding to the recruitment demand information are determined according to the skill similarity to form a candidate set. The method of determining the candidate set according to the skill similarity includes but is not limited to: sorting the candidates according to the skill similarity to select a preset number of candidates; or selecting candidates with a skill similarity greater than the preset to generate a candidate set.

[0128] In practical applications, a large number of similarity calculations can be reduced by determining a candidate set through a clustering algorithm of a knowledge graph. The specific steps in this embodiment are as follows:

[0129] When building a candidate skill knowledge graph, the knowledge graph is clustered into multiple skill communities based on the candidate skill information in the resume. Taking the data analysis skill community as an example, the data analysis skill community contains candidates Liu Yi and Liu Er. The recruitment information for operation recruitment contains skills such as data analysis for operation positions. Based on the skill information in the recruitment information, the target community that matches it, namely the data analysis skill community, is selected, and candidates Liu Yi and Liu Er in the data analysis skill community are selected as the candidate set.

[0130] The candidate determination method of a specific embodiment of the present application generates a job skill knowledge graph and a candidate skill knowledge graph by constructing a knowledge graph, which can normalize skills and effectively reflect the relationship between positions, skills and candidates after normalizing different identifiers of the same skill, which is conducive to obtaining similar positions and similar candidates, solving the problems of low efficiency and strong subjectivity of manual screening, and improving work efficiency. Through the method of similarity calculation, similar positions and candidates can be better screened, and more qualified candidates can be determined from different dimensions for selection by the recruitment demand department, effectively improving work efficiency.

[0131] Corresponding to the above method embodiment, the present application also provides a candidate determination device embodiment based on a knowledge graph, Fig.10 FIG. 1 is a schematic diagram showing a candidate determination device based on a knowledge graph according to an embodiment of the present application. Fig.10 As shown, the device comprises:

[0132] A construction module 1002 is configured to construct a job skill knowledge graph and a candidate skill knowledge graph;

[0133] The generating module 1004 is configured to obtain recruitment demand information and generate a recruitment skill sub-graph based on the recruitment demand information;

[0134] The position determination module 1006 is configured to determine the candidate position corresponding to the recruitment demand information according to the recruitment skill subgraph and the position skill knowledge graph;

[0135] The candidate determination module 1008 is configured to determine a candidate set corresponding to the recruitment requirement information based on the candidate positions and the candidate skill knowledge graph.

[0136] Optionally, the building module 1002 is further configured to:

[0137] Obtain job skills data;

[0138] Parsing the job skill data to obtain job information, skill information corresponding to the job, and job skill weight information;

[0139] A job skill knowledge graph is constructed using the job information and the skill information corresponding to the job as nodes and the job skill weight information as a relationship.

[0140] Optionally, the building module 1002 is further configured to:

[0141] Obtain candidate resumes from a candidate resume database;

[0142] Parsing the candidate resume to obtain candidate information, candidate corresponding skill information and candidate skill weight information;

[0143] The candidate information and the skill information corresponding to the candidate are used as nodes, and the candidate skill weight information is used as a relationship to construct a candidate skill knowledge graph.

[0144] Optionally, the generating module 1004 is further configured to:

[0145] Analyze the recruitment demand information to obtain target position information, skill information corresponding to the target position, and skill weight information of the target position;

[0146] A recruitment skill subgraph is generated with the target position information and the skill information corresponding to the target position as nodes and the target position skill weight information as a relationship.

[0147] Optionally, the position determination module is further configured to:

[0148] According to the graph similarity between the recruitment skill subgraph and the position skill knowledge graph, the candidate position corresponding to the recruitment demand information is determined.

[0149] Optionally, the position determination module 1006 is further configured to:

[0150] Converting the recruitment skill subgraph into a recruitment skill vector, and converting the job skill knowledge graph into multiple job skill vectors;

[0151] Calculating the job similarity between the recruitment skill vector and each job skill vector respectively;

[0152] Determine a candidate position corresponding to the recruitment requirement information according to the position similarity.

[0153] Optionally, the position determination module 1006 is further configured to:

[0154] Sort the positions according to the position similarities, and select a preset number of positions to generate candidate positions as candidate positions corresponding to the recruitment demand information; or

[0155] The positions whose similarity to the positions is greater than a preset threshold are selected as candidate positions corresponding to the recruitment demand information.

[0156] Optionally, the candidate determination module 1008 is further configured to:

[0157] Determining a skill information subgraph corresponding to the candidate position in the candidate skill knowledge graph based on the candidate position;

[0158] According to the skill information subgraph corresponding to the candidate position, a candidate set corresponding to the recruitment demand information is determined in the candidate skill knowledge graph.

[0159] Optionally, the candidate determination module 1008 is further configured to:

[0160] Converting the skill information subgraph corresponding to the candidate position into a candidate position skill vector, and converting the candidate skill knowledge graph into a candidate skill vector;

[0161] Calculating the skill similarity between the candidate position skill vector and the candidate skill vector respectively;

[0162] A candidate set corresponding to the recruitment requirement information is determined according to the skill similarity.

[0163] Optionally, the candidate determination module 1008 is further configured to:

[0164] Sort the candidates according to the skill similarity, and select a preset number of candidates to generate a candidate set; or

[0165] Positions whose skill similarity is greater than a preset threshold are selected to generate a candidate set.

[0166] Optionally, the candidate determination module 1008 is further configured to:

[0167] Cluster candidates according to their skills to generate multiple skill communities;

[0168] Determining a target skill community among the multiple skill communities according to the skill information subgraph;

[0169] The candidates in the target skill community are obtained to generate a candidate set.

[0170] The candidate determination device of a specific embodiment of the present application generates a job skill knowledge graph and a candidate skill knowledge graph by constructing a knowledge graph, which can normalize skills and effectively reflect the relationship between positions, skills and candidates after normalizing different identifiers of the same skill, which is conducive to obtaining similar positions and similar candidates, solving the problems of low efficiency and strong subjectivity of manual screening, and improving work efficiency. Through the method of similarity calculation, similar positions and candidates can be better screened, and more qualified candidates can be determined from different dimensions for selection by the recruitment demand department, effectively improving work efficiency.

[0171] The above is a schematic scheme of a candidate determination device based on a knowledge graph in this embodiment. It should be noted that the technical scheme of the candidate determination device based on the knowledge graph and the technical scheme of the candidate determination method based on the knowledge graph belong to the same concept. For details not described in detail in the technical scheme of the candidate determination device based on the knowledge graph, please refer to the description of the technical scheme of the candidate determination method based on the knowledge graph.

[0172] It should be noted that the components in the device claim should be understood as the functional modules that must be established to implement the steps of the program flow or the steps of the method, and the functional modules are not actually divided or separated. The device claim defined by such a group of functional modules should be understood as the functional module architecture of the solution mainly implemented by the computer program recorded in the specification, and should not be understood as the physical device that mainly implements the solution by hardware.

[0173] In one embodiment of the present application, a computing device is also provided, including a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein the processor implements the steps of the candidate determination method based on the knowledge graph when executing the computer instructions.

[0174] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the candidate determination method based on the knowledge graph are of the same concept, and the details not described in detail in the technical scheme of the computing device can be found in the description of the technical scheme of the candidate determination method based on the knowledge graph.

[0175] An embodiment of the present application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the steps of the candidate determination method based on the knowledge graph as described above.

[0176] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the candidate determination method based on the knowledge graph belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be found in the description of the technical scheme of the candidate determination method based on the knowledge graph.

[0177] An embodiment of the present application discloses a chip storing computer instructions, which, when executed by a processor, implement the steps of the candidate determination method based on the knowledge graph as described above.

[0178] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0179] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0180] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0181] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0182] The preferred embodiments of the present application disclosed above are only used to help explain the present application. The optional embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the present application. The present application selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is only limited by the claims and their full scope and equivalents.

Claims

1. A candidate determination method based on knowledge graph, characterized in that: include: Constructing a job skill knowledge graph and a candidate skill knowledge graph, wherein the job skill knowledge graph is constructed based on the requirements of at least one job, and the candidate skill knowledge graph is constructed based on the information of at least one resume; Acquire recruitment demand information, and generate a recruitment skill subgraph based on the recruitment demand information; Convert the recruitment skill subgraph into a recruitment skill vector, and convert the position skill knowledge graph into multiple position skill vectors; respectively calculate the position similarity between the recruitment skill vector and each position skill vector; determine the candidate position corresponding to the recruitment demand information according to the position similarity, wherein the position similarity between the recruitment skill vector and each position skill vector in the position skill knowledge graph represents the similarity between each skill node in the recruitment skill subgraph and each skill node in the position skill knowledge graph; According to the candidate positions and the candidate skill knowledge graph, a candidate set corresponding to the recruitment demand information is determined.

2. The method according to claim 1, characterized in that: The construction of the job skill knowledge graph includes: Obtain job skills data; Parsing the job skill data to obtain job information, skill information corresponding to the job, and job skill weight information; A job skill knowledge graph is constructed using the job information and the skill information corresponding to the job as nodes and the job skill weight information as a relationship.

3. The method according to claim 1, characterized in that The construction of the candidate skill knowledge graph includes: Obtain candidate resumes from a candidate resume database; Parsing the candidate resume to obtain candidate information, candidate corresponding skill information and candidate skill weight information; The candidate information and the skill information corresponding to the candidate are used as nodes, and the candidate skill weight information is used as a relationship to construct a candidate skill knowledge graph.

4. The method according to claim 1, characterized in that: The generating a recruitment skill subgraph based on the recruitment demand information includes: Analyze the recruitment demand information to obtain target position information, skill information corresponding to the target position, and skill weight information of the target position; A recruitment skill subgraph is generated with the target position information and the skill information corresponding to the target position as nodes and the target position skill weight information as a relationship.

5. The method according to claim 1, characterized in that The determining of the candidate positions corresponding to the recruitment demand information according to the position similarity includes: Sort the positions according to the position similarities, and select a preset number of positions to generate candidate positions as candidate positions corresponding to the recruitment demand information; or The positions whose similarity to the positions is greater than a preset threshold are selected as candidate positions corresponding to the recruitment demand information.

6. The method according to claim 1, characterized in that The step of determining a candidate set corresponding to the recruitment demand information according to the candidate positions and the candidate skill knowledge graph includes: Determining a skill information subgraph corresponding to the candidate position in the candidate skill knowledge graph based on the candidate position; According to the skill information subgraph corresponding to the candidate position, a candidate set corresponding to the recruitment demand information is determined in the candidate skill knowledge graph.

7. The method according to claim 6, characterized in that The determining, in the candidate skill knowledge graph according to the skill information subgraph corresponding to the candidate position, a candidate set corresponding to the recruitment demand information includes: Converting the skill information subgraph corresponding to the candidate position into a candidate position skill vector, and converting the candidate skill knowledge graph into a candidate skill vector; Calculating the skill similarity between the candidate position skill vector and the candidate skill vector respectively; A candidate set corresponding to the recruitment requirement information is determined according to the skill similarity.

8. The method according to claim 7, characterized in that The determining of a candidate set corresponding to the recruitment requirement information according to the skill similarity includes: Sort the candidates according to the skill similarity, and select a preset number of candidates to generate a candidate set; or The candidates whose skill similarity is greater than a preset threshold are selected to generate a candidate set.

9. The method according to claim 6, characterized in that The determining, in the candidate skill knowledge graph according to the skill information subgraph corresponding to the candidate position, a candidate set corresponding to the recruitment demand information includes: Cluster candidates according to their skills to generate multiple skill communities; Determining a target skill community among the multiple skill communities according to the skill information subgraph; The candidates in the target skill community are obtained to generate a candidate set.

10. A candidate determination device based on knowledge graph, characterized in that: include: A construction module is configured to construct a job skill knowledge graph and a candidate skill knowledge graph, wherein the job skill knowledge graph is constructed based on the requirements of at least one job, and the candidate skill knowledge graph is constructed based on the information of at least one resume; A generating module is configured to obtain recruitment demand information and generate a recruitment skill subgraph based on the recruitment demand information; The position determination module is configured to convert the recruitment skill subgraph into a recruitment skill vector, and convert the position skill knowledge graph into a plurality of position skill vectors; respectively calculate the position similarity between the recruitment skill vector and each position skill vector; and determine the candidate position corresponding to the recruitment demand information according to the position similarity, wherein the position similarity between the recruitment skill vector and each position skill vector in the position skill knowledge graph represents the similarity between each skill node in the recruitment skill subgraph and each skill node in the position skill knowledge graph; The candidate determination module is configured to determine a candidate set corresponding to the recruitment demand information based on the candidate positions and the candidate skill knowledge graph.

11. A computing device comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, characterized in that: When the processor executes the computer instructions, the steps of the method according to any one of claims 1 to 9 are implemented.

12. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method described in any one of claims 1 to 9 are implemented.

Images