Text Classification Algorithm And Its Realization In The Campus Recruitment Management System

With the development of office automation, the request for paperless management which replaces the manual management approach to be growing. Some large companies have set up their own recruitment system which allowed job hunters submit their resumes through its electronic system. With the increase of the company's personnel needs and the number of job seekers, electronic resume management and flitting becomes a hard wok for recruiters. In order to enhance the resume management and automated fliting work, this thesis designs and implements campus recruitment management system based on a text classification algorithm. And this system has been applied in the company's recruitment.The main work of this thesis is as follows:1. Studied IG feature selection algorithm and improved some part of this algorithm. Because the IG algorithm prefers low-frequency words, so we combined frequency, degree of dispersion with IG feature selection algorithm, in order to make up for this disadvantage, and this method has been proven to be effective. Studied CHI feature selection algorithm and the improved strategies for CHI algorithms; and verified the improvement of the performance of CHI algorithm through experiments through experiment. And compared the performance of the improved IG feature selection algorithm and improved CHI feature selection algorithm.2. Studied and realized C4.5 decision tree algorithm and SVM classification algorithm. Studied the information gain ratio and pessimistic error pruning of C4.5 decision tree, and carried out the classification results of C4.5 decision tree algorithm by Weka software. Studied SVM classification algorithm principle, kernel function and penalty parameter, and determined the optimal value of this three parameters by experimental Analysis.3. Designed the of campus recruitment management system based on the improved text classification. The thesis designed the system function architecture, development architecture and development framework, and focus on the design of the 3 innovative modules: student resume management module, resume matching module and assessment Star management module, which combined text classification algorithm.4. Realized student resume management module, resume matching module and assessment Star management module. Realized the function of student resume Excel exporting, SVM classifier predictions Excel importing, predictions match and the assessment of star standard based on C4.5 decision tree results.Run the system for each module, which shows the campus recruitment management system is stable, easy to operate. By adding the optimized SVM classifier prediction with the campus recruitment management system, the system got a better performance on the ability of guidance.