Feature-based Sentiment Analysis on Android App Reviews Using SAS® Text Miner and SAS® Sentiment Analysis Studio
Jiawen Liu, Mantosh Kumar Sarkar, Goutam Chakraborty · 2013
Sentiment analysis is a popular technique for summarizing and analyzing consumers ’ textual reviews about products and services. There are two major approaches for performing sentiment analysis; statistical model based approaches and Natural Language Processing (NLP) based approaches to create rules. In this study, we first apply text mining to summarize users ’ reviews of Android Apps and extract features of the apps mentioned in the reviews. We then use NLP approach for writing rules. We use reviews of two recent apps; a widget app from Brain & Puzzle category and a game app from Personalization category. We extracted six hundred textual reviews for each app from Google Play Android App Store. SAS ® Enterprise Miner TM 7.1 is used for summarizing reviews and pulling out features, and SAS ® Sentiment Analysis Studio 12.1 is used for performing sentiment analysis. Our results show that for both apps, carefully designed NLP rule-based models outperform the default statistical models in SAS ® Sentiment Analysis Studio 12.1 for predicting sentiments in test data. NLP rule based models also provide deeper insights than statistical models in understanding consumers ’ sentiments.