Construction And Application Of Knowledge Graph Matching Talent Training In Vocational Colleges And Job Demand

Vocational education is oriented by vocational demand,and the matching degree between talent training and job demand is an important index to measure the quality of vocational education.With the development of social economy and the adjustment of industrial structure,the updating and iteration speed of posts is also accelerating.Vocational colleges and universities take talent training plan as the training criterion.When studying the matching degree between talent training and job demand,it is necessary to timely and accurately understand the change of job demand,so as to provide a reasonable basis for the adjustment of training plan.Recruitment data is the most intuitive mapping of job demand changes.At present,with the development of the Internet,a large amount of information rich but complicated network recruitment data has been produced.As a representation of knowledge,knowledge graph can show existing connections between data and discover deeper connections.By constructing the knowledge graph based on training programs and recruitment data,we hope to explore the correlation between the skills developed by colleges and those required in recruitment information,so as to provide data support for the dynamic adjustment of training programs in vocational colleges.The main work of this paper is as follows:Data acquisition and preprocessing.In this paper,the training plan and recruitment information of computer-related majors in vocational colleges are taken as examples.Among them,the data of course name and curriculum setting in the training plan are relatively standard,and they are converted into subsequent corpus through file format.Online recruitment information is mostly unstructured text information.Python is used as the development tool to access 51job.com recruitment website.Data are preprocessed through regular expression and word segmentation,and the word cloud tool is used to visually display the composition of the data.Training positions demand entity recognition model based on BiLSTM-CRF,extracted from the recruitment information skills name,master degree,4 kinds of computer industry work experience,personality requirements for entities,in the heart of the identification process data sequence annotation using BIO system,through the word embedding algorithm into low-dimensional dense term vectors as the input of neural network BiLSTM layer to complete the feature extraction output per word labels,on top of this,increased the CRF layer between the label constraints to improve the effect of recruitment entity recognition,in the data set was determined by the multiple sets of contrast experiment on the performance of optimal parameters setting.In this paper,a knowledge graph construction method was proposed to match talent training programs with occupational demands,and a prototype system was designed and implemented based on the knowledge graph.Based on the data of training program,the knowledge map of curriculum setting of computer specialty is constructed.Based on the online recruitment information,the paper constructs the job demand knowledge map.The fusion of the graph is accomplished by establishing the association of skill entities.The concrete construction process includes four steps:entity extraction,relationship definition,knowledge storage and graph fusion.The prototype system of knowledge graph display query is designed and implemented using Django framework.The data is stored in the Neo4j graph database,which can identify the entity of the input text information through the system,and provide such functions as entity query and graph relationship display.