Title:
SENTENCE VECTOR CALCULATION METHOD BASED ON CHI-SQUARE TEST, AND TEXT CLASSIFICATION METHOD AND SYSTEM
Document Type and Number:
WIPO Patent Application WO/2020/063071
Kind Code:
A1
Abstract:
Disclosed are a sentence vector calculation method based on a chi-square test, and a text classification method and system, the method involving: carrying out word segmentation processing on the current text and removing stop words to obtain a word segmentation result; calculating a word vector of each word in the word segmentation result; calculating a chi-square value between each word vector and a preset category, and dividing the word vectors into feature words and non-feature words according to the chi-square values; calculating usage frequency of the feature words in the preset category, giving a first weight value to the feature words according to the usage frequency, and giving a second weight value to the non-feature words, wherein the first weight value is greater than the second weight value; and calculating a weighted mean value of all word vectors according to the word vectors of the feature words and the non-feature words and the corresponding weight values, and taking same as a sentence vector of the current text, thereby improving a weight value of the sentence vector in a feature dimension, reducing mutual interference between the word vectors in text information and greatly improving the accuracy of text classification.
Inventors:
HUANG YOUFU (CN)
XIAO LONGYUAN (CN)
CAI ZHENHUA (CN)
LI XIMIN (CN)
LIU XIAOWEI (CN)
TAN YUKUN (CN)
XIAO LONGYUAN (CN)
CAI ZHENHUA (CN)
LI XIMIN (CN)
LIU XIAOWEI (CN)
TAN YUKUN (CN)
Application Number:
PCT/CN2019/097187
Publication Date:
April 02, 2020
Filing Date:
July 23, 2019
Export Citation:
Assignee:
XIAMEN KUAISHANGTONG INFORMATION TECH CO LTD (CN)
International Classes:
G06F17/27
Domestic Patent References:
| WO2014002775A1 | 2014-01-03 |
Foreign References:
| CN109522544A | 2019-03-26 | |||
| CN107229610A | 2017-10-03 | |||
| CN107818077A | 2018-03-20 | |||
| CN107622333A | 2018-01-23 |
Attorney, Agent or Firm:
XIAMEN SHICHENG ASSOCIATES INTELLECTUAL PROPERTY AGENCY (CN)
Download PDF: