Optimization method and device for text categorization model

Abstract

The invention relates to the technical field of computers, in particular to an optimization method and device for a text categorization model. The optimization method for the text categorization modelcomprises the steps that a text set is firstly acquired, preprocessing is carried out on each text in the text set, and a term set corresponding to each text is obtained. Terms in the term set are matched with various category feature words in a preset feature word set, according to a matching result, a term category in the term set is determined. According to the term category, the preset feature word set is expanded. According to the expanded preset feature word set, terms in the tern set are filtered. Weighted values of the filtered terms in the term set are determined, and according to the filtered terms in the term set and the corresponding weighted values, the optimization is carried out on a preset text categorization model. Therefore, the text categorization model which can accurately categorize texts can be obtained.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a text classification model optimization method and device. Background technique [0002] In traditional technology, texts are usually classified by the following two methods: [0003] The first method is a rule-based method, which collects some commonly used category keywords in advance, and the categories corresponding to these category keywords are known. When the text to be classified matches a certain category keyword, the The text is classified into the category corresponding to the category keyword. However, this method has great limitations. When a text does not match any category keywords, the text cannot be classified. Furthermore, this method usually cannot accurately classify text. For example, assume that the pre-collected category keyword is: "Tmall", and its corresponding category is the Internet industry. Since the text "Tmall Service Station" cont...

AI Technical Summary

Problems solved by technology

However, this method is usually only suitable for classifying long texts, because in long texts, some words will appear multiple times, that is, TF can play a corresponding role

For short text (the text usually only includes a few words), most of the words appear only once in the text, TF has no meaning, so according to the above text classification model, the short text cannot be accurately classified

Images