arrow
返回

A study for extracting keywords from data with deep learning and suffix array

delete2022-01-26
delete1
PRE
AI
W
Wentao Xu
DOI:10.1007/s11042-021-11762-7delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
While a suffix index built on a suffix array is capable of supporting full-text searches over any data, its search speed can be accelerated using a keyword index for the set of keywords extracted from data. We attempt to design a method for extracting keywords from data using deep learning and a suffix array in this article. In particular, the study starts with Chinese texts because many word segmentation results on Chinese are available for performance evaluation. A new method combining the use of a neural network and a suffix array of training data is proposed for Chinese word segmentation. The suffix array of training data is constructed to divide long sentences in the input text into short fragments for better word segmentation by our neural network method without a context window. Our experiments on the typical datasets reveal that the proposed method achieves encouraging results in terms of the precision, recall and F-1 score compared to other existing advanced methods while avoiding the drawback of a context window. This study provides some helpful experience for designing a general solution to extract keywords from data using a suffix array.
Keyword:
Extracting keywords
Chinese word segmentation
Neural network
Suffix array

期刊

Multimedia Tools and Applications 封面图
Multimedia Tools and Applications
IF:
3
论文数:
2.0W
被引数:
3.2W

机构

S
Sun Yat Sen University
学者数:
9.9W
论文数: 7.2W
被引数: 95
引用论文

引用论文

Thermodynamics study on the generation of electricity via CO2-mineralization cell
err2015-08-23
err0
errOAAI
errHeping Xie; Wen Jiang; Yufei Wang; Tao Liu; Ruilin Wang; Bin Liang; Yang He; Jinlong Wang; Liang Tang; Jinwei Chen
err分享
err收藏
Domain phrase identification using atomic word formation in Chinese text
err2011-12-01
err6
PREAI
errLiu, Qingtang; Wu, Linjing; Yang, Zongkai; Liu, Yaoyao
err分享
err收藏
学者 查看更多内容