Research On The Measurement Of Person Post Matching Of Internet Technicians Based On CNN

With the rapid development of Internet technology,the rise of network recruitment enables the rapid spread of talent information in unlimited regions.In order to adapt to the changes of the media in the talent market,enterprises are gradually promoting the development of electronic human resource management.Especially in the Internet enterprises with technology as the core,the computer recruitment system has been able to realize the intelligent processing of part of the recruitment work,so the traditional "experience + intuition" mode of person post matching is no longer applicable,and the intelligent measurement of person post matching has become the new goal of the electronic human resource management of Internet enterprises.In the background of big data,machine learning algorithm enables the computer to obtain information automatically,so as to make scientific and effective prediction.It is an inevitable trend of human resource management electronization to apply big data technology to human post matching measurement scenarios,improve the efficiency of human post matching through effective machine learning methods,and reduce the cost ratio.In order to explore the performance of convolutional neural network algorithm with excellent performance in human post matching measurement in various scenes such as image recognition,natural language processing,etc.,this paper first uses the web crawler technology to crawl the real post demand data of 2775 pull hooks of Internet enterprises,constructs the index pool through high-frequency word statistics,and extracts the index based on expert interview suggestions To the evaluation index system,which includes three dimensions of knowledge level and talent quality,and 11 characteristic indexes.On this basis,this paper realizes the transformation of talent evaluation data from vector to 16 order square matrix by building a fully connected neural network feature converter,and embeds it into the structure of convolution neural network classifier,forming a person position matching measurement model based on convolution neural network.Finally,this paper downloads 19832 resumes of technical position talents through the pull hook network,extracts empirical data by using the evaluation index system obtained from the research,realizes model fitting through python programming,eliminates the contingency of fitting results by combining different number of sample division,and carries out volume neural network model and decision tree,naive Bayes two traditional machine learning multi classification algorithms The comparative analysis of the prediction results,the influence of the learning rate of convolutional neural network,the number of all connected layers and the optimization algorithm on the prediction results,and the application of the model in the actual recruitment of enterprises.The empirical research shows that the loss of the model is reduced to 0.413 after 300 iterations,and the accuracy is as high as 0.844,which can be used to solve the practical problems of the human post matching measurement.Compared with the accuracy of decision tree 0.738 and naive Bayes 0.624,the convolution neural network model has better prediction effect.The single full connection convolution model with random gradient descent and 0.001 learning rate has a shorter calculation time while maintaining a certain accuracy,which is more in line with the actual application scenarios of enterprises.In practical application,using the convolution model of human post matching measure proposed in this paper to calculate the score of talent matching will save nearly 90% of the time cost,and the effect is basically the same as that of human resource matching.