多分类器融合的短文本识别算法  

Short text recognition algorithm based on multi-classifier fusion

作　　者：杨妍 赵军民[1] 

机构地区：[1]河南城建学院,河南平顶山450001

出　　处：《信息通信》2019年第3期22-23,共2页Information & Communications

基　　金：国家语委语言文字科研项目优秀成果后期资助计划"情感词语知识库;语料库建设及应用"(项目编号:HQ135-16)

摘　　要：近年来,关于短文本的文本挖掘逐渐变成了自然语言处理的热点之一。针对微博这种短文本,内容较少但语义却极其丰富,而且新词、错字较多等特征,文章提出一种多分类器融合的短文本识别算法,该算法能够自动的对短文本情感倾向性进行识别。从NLPCC2012提供的评测数据的实验结果来看,该方法取得了一定的效果。In recent years,text mining on short texts has gradually become one of the hot topic of natural language processing.For the short text ofMicro-blog,the content is small,but the semantics are rich,and the new words and typos are more.This paper proposes a short text recognition algorithm based on multi-classifier fusion,which can automatically match short texts.Emotional orientation is identified.From the experimental results of the evaluation data provided by NLPCC2012,the method has achieved certain results.

关 键 词：情感分析 情感词 分类器 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]