Design And Implementation Of Talent Evaluation System Based On Name Entity Recognition And Semantic Similarity Calculation

For the country,development requires the support of talents,and training talents requires a lot of time and resources.In order to save the cost of talent training,the Chinese government has proposed a series of talent introduction policies.Therefore,how to establish a scientific,fair and efficient talent evaluation system based on these policies has become very important.At present,the talent evaluation systems established in various regions of my country are too subjective,the evaluation efficiency is low,and it is impossible to effectively classify talent information.In order to improve the efficiency and accuracy of talent evaluation,this paper improves the existing talent information analysis and evaluation method based on named entity recognition and semantic similarity calculation technology,and builds a talent evaluation system combined with the microservice framework to realize automatic evaluation of talent information.The main work of this paper is as follows:(1)By sorting out the talent introduction policies of the country and the Zhejiang Provincial Government,combined with the national senior engineer qualification quantitative evaluation standard to evaluate and quantify the talent information,a set of talent type and score evaluation model was designed;combined with the front-end and back-end separation The service framework designs the software solution of the system.(2)Aiming at the problem that the amount of talent information is large and the text sequence is too long,which makes it difficult to evaluate the talent information,the named entity recognition technology that integrates the two-way long-short memory network and the conditional random field is used to identify and extract the text data of the talent,and according to the talent The type evaluation process divides the talent data into nine types of entity tags,and then realizes the classification of talent information.(3)To solve the problem of inaccurate matching of talent and score labels,by analyzing the text features of talent labels and score labels,using the semantic similarity calculation method based on deep learning,adding Bi LSTM network and attention mechanism to improve the Bert model,designing and realized the talent tag score calculation module.(4)Build a talent evaluation system,divide the system modules through the talent evaluation process and talent categories,adopt the micro-service architecture,and use the Mysql+Neo4j+Mongo DB database to design and complete the talent evaluation system to achieve system testing and verification.The research results show that the talent evaluation quantitative scheme proposed in this paper reduces the difficulty of talent evaluation and improves the efficiency and accuracy of talent evaluation;the design and implementation of the talent evaluation system using the micro-service architecture with front-end and back-end separation can meet the system scalability and easy maintenance;in addition,the named entity recognition model that integrates two-way long-short memory network and conditional random field realizes the preprocessing of talent information and improves the speed of talent evaluation;the improved Bert model solves the problem of Bert model words The uneven distribution of vectors and the inability to focus on key semantic information improve the accuracy of text matching.