{"title":"Using genetic algorithms in word-vector optimisation","authors":"P. Smith","doi":"10.1109/UKCI.2010.5625589","DOIUrl":null,"url":null,"abstract":"Word vectors and sets of words are used in a wide range of text-based applications. Yet these word sets are often chosen on an ad hoc basis. In this study, we examine two text-based applications that use word sets and in both cases find that classification performance can be optimised using a fairly simple genetic algorithm. The first study is in authorship attribution, the second one is sentiment analysis and in both cases classification precision can be improved using a genetic algorithm. In authorship attribution, in recent years the trend has been towards ever larger word vectors [1,2]. We suggest that this might be a counter-productive step as it can easily lead to inaccuracy caused by overfitting or vector-space sparsity (the curse of dimensionality). In sentiment analysis precision is the main issue as rates of greater than 80–85% are not easy to achieve.","PeriodicalId":403291,"journal":{"name":"2010 UK Workshop on Computational Intelligence (UKCI)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2010-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2010 UK Workshop on Computational Intelligence (UKCI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/UKCI.2010.5625589","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Word vectors and sets of words are used in a wide range of text-based applications. Yet these word sets are often chosen on an ad hoc basis. In this study, we examine two text-based applications that use word sets and in both cases find that classification performance can be optimised using a fairly simple genetic algorithm. The first study is in authorship attribution, the second one is sentiment analysis and in both cases classification precision can be improved using a genetic algorithm. In authorship attribution, in recent years the trend has been towards ever larger word vectors [1,2]. We suggest that this might be a counter-productive step as it can easily lead to inaccuracy caused by overfitting or vector-space sparsity (the curse of dimensionality). In sentiment analysis precision is the main issue as rates of greater than 80–85% are not easy to achieve.