{"title":"Extracting Software Change Requests from Mobile App Reviews","authors":"Muhammad Nadeem, Khurram Shahzad, N. Majeed","doi":"10.1109/ASEW52652.2021.00047","DOIUrl":null,"url":null,"abstract":"The use of mobile apps is increasing rapidly. These apps have thousands of reviews which are widely acknowledged as a valuable resource for the community involved in the development of mobile apps. In this study, we contend that these reviews can be used to generate software change request document for improving mobile apps. A pre-requisite for generating such a document is the identification of Software Change Requests (SCR) from the user reviews. However, the manual processing of these large number of reviews to identify SCRs is a resource intensive task. However, most of the existing studies have focused on the identification of bugs. Whereas, a few studies have been conducted to identify change requests and its localization from mobile apps review, which substantially different from extracting SCR. To that end, we have scrapped review of seven Mobile Apps and developed a dataset that can be used for training of machine learning techniques for the automatic identification of SCRs. A key feature of the approach is that we have documented the annotation guidelines that are used to distinguish between SCR and non-SCR sentences. These guidelines can be used to enhance the developed dataset, as well as to develop new datasets. As another contribution, we have evaluated the effectiveness of five supervised learning techniques for their ability to identify SCR sentences from user reviews. The study shows that Logistic Regression achieved a nearly perfect F1 score of 0.97 for extracting SCR from textual reviews.","PeriodicalId":349977,"journal":{"name":"2021 36th IEEE/ACM International Conference on Automated Software Engineering Workshops (ASEW)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 36th IEEE/ACM International Conference on Automated Software Engineering Workshops (ASEW)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ASEW52652.2021.00047","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
The use of mobile apps is increasing rapidly. These apps have thousands of reviews which are widely acknowledged as a valuable resource for the community involved in the development of mobile apps. In this study, we contend that these reviews can be used to generate software change request document for improving mobile apps. A pre-requisite for generating such a document is the identification of Software Change Requests (SCR) from the user reviews. However, the manual processing of these large number of reviews to identify SCRs is a resource intensive task. However, most of the existing studies have focused on the identification of bugs. Whereas, a few studies have been conducted to identify change requests and its localization from mobile apps review, which substantially different from extracting SCR. To that end, we have scrapped review of seven Mobile Apps and developed a dataset that can be used for training of machine learning techniques for the automatic identification of SCRs. A key feature of the approach is that we have documented the annotation guidelines that are used to distinguish between SCR and non-SCR sentences. These guidelines can be used to enhance the developed dataset, as well as to develop new datasets. As another contribution, we have evaluated the effectiveness of five supervised learning techniques for their ability to identify SCR sentences from user reviews. The study shows that Logistic Regression achieved a nearly perfect F1 score of 0.97 for extracting SCR from textual reviews.