OSPAci: Online Sentiment-Preference Analysis of User Reviews for Continues App Improvement

摘要

Detecting user's sentiment and preference (e.g., complain or new feature wanted) timely and precisely is crucial for developers to improve their apps correspondingly to win the competitive mobile-app market. In this paper, we propose a novel and automated framework OSPAci, which aims to identify user's sentiment and preference effectively based on online user reviews. OSPAci uses sentiment analysis and natural language processing techniques to obtain sentence-level sentiment scores and fine-grained user preference from mobile app reviews. Then, it analysis the evolution of user's sentiment trend and preference. Finally, the user sentiment trend and preference correlation is analyzed along the time dimension, thus this model can be used to monitor user's sentiment tendency and preference almost in time. We evaluate the feasibility and performance of OSPAci by using real Google play's user reviews. The experimental results show that OSPAci can effectively and efficiently identify the user's sentiment tendency and detect user preference timely and precisely.