摘要
随着智能手机的普及,APP软件越来越流行,随之而来的是APP软件用户评论的增多。在数量极大的评论中,关于APP软件缺陷问题的评论是APP开发者最关心的。通过对APP评论的大量阅读和观察,发现APP软件缺陷问题是分散的。总结了7类缺陷问题,使用改进卡方统计和APP软件简介中的名词和动词作为特征选择思路,使用朴素贝叶斯算法对每个缺陷问题评论进行训练学习。用8 677条评论进行实验,结果表明该方法的准确率、召回率和F1值较高。该方法不仅减轻了人工标记APP缺陷问题评论的工作量,而且提高了分类准确度。
With the popularity of smart phones,APP software is becoming more and more popular,followed by the increase of APP software users＇ comments.In a large number of comments, the comments on APP software defects are the core issues APP developers most concern.For the defects of APP software are scattered,7 kinds of defects are summarized,and then the improved Chi square statistics and the nouns and verbs of the introduction of APP software employed as feature selection ideas,and training study of the defects of each comment is conducted by Naive Bayesian algorithm.8677 comments are taken in the experiment and the experimental results show that the accuracy,recall and F1 value of the method are high.It is concluded that this method not only reduces the workload of the comment on APP defects,but also improves the accuracy of the classification.
作者
王延飞
WANG Yan-fei(Department of Information Engineering and Automation,KunmingUniversity of Science and Technology,Kunming 650500,China)
出处
《软件导刊》
2018年第9期59-63,共5页
Software Guide
