Abstract | ||
---|---|---|
User-generated mobile application reviews have become a gold mine for timely identifying functional defects in this type of software artifacts. In this work, we develop a hidden structural SVM model for extracting detailed defect descriptions from user reviews at the sentence level. Structured features and constraints are introduced to reduce the demand of exhaustive manual annotation at the sentence level and enable the use of partially annotated review data for model training. Extensive empirical evaluations on a large collection of mobile application reviews collected from Apple App Store demonstrate the effectiveness of our proposed solution in recognizing the user-reported implementation defects from review content, especially when only partial annotation is available. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/ICDMW.2017.61 | 2017 IEEE International Conference on Data Mining Workshops (ICDMW) |
Keywords | Field | DocType |
user-reported mobile application defects,online reviews,mobile application reviews,software artifacts,user-reported implementation defects,review content,functional defects,structural SVM model,detailed defect description extraction,Apple App Store,mobile app reviews | Data modeling,Annotation,Information retrieval,App store,Computer science,Support vector machine,Feature extraction,Software,Artificial intelligence,Sentence,Mobile telephony,Machine learning | Conference |
ISSN | ISBN | Citations |
2375-9232 | 978-1-5386-3801-9 | 1 |
PageRank | References | Authors |
0.34 | 16 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yue Wang | 1 | 960 | 143.63 |
Hongning Wang | 2 | 925 | 54.89 |
Hui Fang | 3 | 918 | 63.03 |