A new analysis method for user reviews of mobile fitness apps

Abstract

Keeping fit has always been the focus of public attention, so the emergence of fitness apps has been highly sought after. However, after the outbreak, updates to fitness apps often disappoint their users. Therefore, it is eager for developers to learn about the real requirements and opinions to their apps. To investigate users’ requirements, we take data mining on user reviews, and apply sentiment analysis to obtain users’ evaluations and suggestions on the attributes and functions of fitness apps. The users of fitness apps have a high follow-up, and usually have been using these apps for a long time, leaning to treat them as daily applications, and can clearly perceive the experience brought by apps during the process. Therefore, users are very serious when evaluating the apps, and the reviews include user requirements, ideas for improvements, positive and negative sentiments about specific features, and descriptions of experiences with these features. Based on the characteristics of the reviews, we use the P-N deep analysis method to perform sentiment analysis on user reviews. For each review, we extract the active and negative evaluations of the corresponding features separately to avoid errors in evaluating features of apps only based on the star rating of the reviews. And when extracting sentiment words in sentiment analysis, verbs are added as features words and adverbs are added as emotion words because users of fitness apps use more adverbs than that of other apps to express their feelings and many nouns can be expressed by verbs in Chinese.