面向领域新闻的词汇输入预测  

Lexical input prediction for domain news

作　　者：张明西[1] 马悦荣 林启新 ZHANG Ming-xi;MA Yue-rong;LIN Qi-xin(College of Communication and Art Design,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学出版印刷与艺术设计学院,上海200093

出　　处：《计算机工程与设计》2023年第1期262-268,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(62002225);上海市自然科学基金项目(21ZR1445400)。

摘　　要：为解决领域新闻中词汇使用频率对输入预测准确度的影响,提出一种面向领域新闻的词汇输入预测方法,结合样本修剪提升LSTM应用于领域新闻词汇预测中的准确性。利用文本的词序关系形成词汇网络,构建词汇相关度度量模型计算词汇间的相关度,依据词汇间的强弱关系进行词汇修剪。基于词序关系,采用LSTM进行训练,生成词汇输入预测序列。实验结果表明,相比传统LSTM模型,所提方法能够提升平均4.73%的准确率,验证了所提方法的可行性与有效性。To solve the influence of word frequency on the accuracy of input prediction,a word input prediction method for domain news was proposed,which combined with sample pruning to improve the accuracy of LSTM in the prediction of domain news words.The word order relation of text was used to form a vocabulary network,and a measure model of vocabulary relatedness was constructed to calculate the relatedness among words.Based on the word order relation,LSTM was used to train the word input prediction sequence.Experimental results show that compared with the traditional LSTM model,the proposed method can improve the accuracy by 4.73%on average,which verifies the feasibility and effectiveness of the proposed method.

关 键 词：长短期记忆网络 领域新闻 词汇输入预测 词汇网络 TFIDF模型 

分 类 号：TP317.2[自动化与计算机技术—计算机软件与理论]