SVM text classification-based accurate resume parsing method

Abstract

The invention relates to an SVM text classification-based accurate resume parsing method comprising the following steps: (1) Microsoft office is operated under net framework, resume files in all kinds of formats are converted into a PDF format, and then the files are converted into an xml format from the PDF format; (2) labels of all resume text lines in the xml format are extracted and corresponding feature vectors are generated; (3) all the resume text lines are marked, an SVM approach is adopted for classification based training according to marked values and the feature vectors corresponding to all the resume text lines, and a classifier is obtained; (4) all resumes are cut according to the obtained classifier, extracted information is parsed in a partitioned manner, and therefore all the resumes can be parsed accurately.

Description

technical field [0001] The invention relates to a resume accurate analysis method based on SVM text classification, which is natural language processing, pattern recognition, AC automaton search technology and .net operation Microsoft word technology, and is a resume accurate analysis method integrating multiple technologies. Background technique [0002] At present, the general method of uploading resume analysis solutions on human resource websites in the market is as follows: convert the files uploaded by users into plain text format, list the dictionary of field names that need to be parsed, and then look up the words in these dictionaries in the resume. The word will return the content within a certain range in the future as the analysis result of this field. There are generally two big disadvantages in doing this: first, if a resume keyword appears in the middle of a large piece of text, it will disconnect the original content, and the analysis result of this method wi...

AI Technical Summary

Problems solved by technology

There are generally two big disadvantages in doing this: first, if a resume keyword appears in the middle of a large piece of text, it will disconnect the original content, and the analysis result of this method will be It will be completely wrong, resulting in the lack of information in the analysis results, and the robustness of the algorithm is poor.

Second, if all files are converted into text when uploading, the original resume format information will be lost, and correspondingly a lot of useful content information will be lost

Images