Application Of Improved Na?ve Bayes Algorithm And KNN Algorithm In Recruitment Text Categorization

With the rapid development of the Internet era,vast amounts of data are available over the Internet.Likewise,job seekers can also get job information from job websites on the Internet.But there is some confusion on these sites: there are lots of irrelevant ads,for example,or job postings with vague needs.The occurrence of these situations,for job seekers looking for added a lot of trouble on the way to work,not only need to waste a lot of time to browse the invalid recruitment information,also need to spend effort to analyze whether the recruitment information correct "recruitment information",the emergence of these circumstances,would greatly influence the efficiency of the applicant to apply for a job.In order to improve this situation,the recruitment text can be classified and processed.If the job information is accurately classified,the text that does not belong to the job information can be removed.In that case,job seekers only need to browse the job information they are interested in in the text that belongs to the job information.Usually,the algorithms used for text classification are: Na?ve Bayes algorithm、 KNN algorithm、 support vector machine etc.According to the feature of recruitment information,this paper changes the calculation methods of Na?ve Bayes algorithm and KNN algorithm.,and realized the improved algorithm to complete the work of more accurate classification of recruitment text.The main work of this paper is:(1)Improved Na?ve Bayes algorithm.A no-zero Na?ve Bayes(NZ-NB)algorithm was proposed and validated.This paper proposes a Na?ve Bayes text classification algorithm for recruitment text classification.By analyzing the characteristics of the recruitment text and Na?ve Bayes algorithm,it is found that the Na?ve Bayes algorithm has the problem of zero probability,and it is improved to screen the data set and select the recruitment text and the non-recruitment text.By comparing the results of NZ-NB algorithm and the original Na?ve Bayes algorithm,it is verified that the classification accuracy of NZ-NB algorithm is relatively good.(2)Improved KNN classification algorithm,proposed FW-KNN(Fast-Weight KNN)algorithm,and its verification.For the original KNN algorithm in the face of recruitment text,classification efficiency is low.The original KNN algorithm is improved,and the professional terms existing in the recruitment text are weighted,so that the characteristic words with the same weight are calculated only once,and the classification efficiency is improved.By comparing the results of the original KNN algorithm and FW-KNN algorithm,it is verified that the improved algorithm has faster efficiency.The NZ-NB algorithm and FW-KNN algorithm proposed in this paper have better classification accuracy and classification efficiency compared with the previous improved Na?ve Bayes algorithm and KNN algorithm.