A method on Chinese thesauri

Abstract

In recent years, text analysis has become increasingly heated in many fields. And now, majority methods of text analysis are using Word2vec, Naïve Bayes or so on to classify the large number of texts. But for the text itself, not all samples are useful for some high-requirement researches and only use one keywords to get the related sample is definitely not enough. In this paper, we provide a novel model of second text filtering with Chinese Thesauri. It includes roughly 5 steps: sample collecting, thesauri establishment, word-segment algorithm, word-frequency statistics and the calculation of text relevance. Its main purpose is making the sample texts more accurate with the keywords which are input by the user and avoiding the needless time and space waste.