Abstract
Play Store reviews play an important role in demonstrating that decisions are made from the user’s perspective, and contain a wealth of knowledge that can be used to understand quality issues and help developers build higher-quality mobile applications. Even for very important information, it can ensure the authenticity of user-generated content. In Play Store, wearable applications were recently launched, and are always open and easy to use, and are gradually being welcomed by users. Driven by popularity and self-interest, profit-incentive developers are developing low-quality applications and hiring robots to exaggerate ratings, reviews, or install counts. This is how spam in applications increases. Low-quality applications reduce the user’s quality experience and trust, because after users download an application, they will know the irrelevant and annoying content of the application. As a result, the reputation of the Play Store is damaged. Therefore, we analyzed the review content of different wearable applications and proposed a regression model that has a wide range of recommended features, including sentiment, content similarity, language and time features, to detect wearable applications in the Play Store. We use advanced machine learning techniques to evaluate and verify the quality of the model. Compared with existing models, the performance of our proposed model is very good, with an error rate as low as 0.40 MSE. Therefore, our regression model is most suitable for deep neural network (DNN) training.
Highlights
From 2009 to 2019, the development of mobile applications on the market increased exponentially, with approximately 3.3 million available applications recorded [1]
It can be observed that the deep neural network (DNN) model performs well compared with all implemented regression models, with a smaller mean absolute error (MAE) value of 3.1
Our regression model is developed through feature engineering, involving multiple features, such as emotion, language, content similarity, and temporal features
Summary
From 2009 to 2019, the development of mobile applications on the market increased exponentially, with approximately 3.3 million available applications recorded [1]. Liu et al [10] uses text mining to extract features of applications and summarize user comments related to different applications and Natural Language Processing (NLP) techniques for writing rules. In Guzman et al [12] uses NLP technology to extract application features in user reviews and topic modeling in order to group these features into more meaningful features By combining these two technologies, the author generates abstracts with different granularities, which will help analyze user comments to identify new user needs or schedule future versions. Therein, the authors use app review ratings, package names, and main conversations to detect fraud in apps They apply NLP technology to obtain action words and fuzzy logic to further classify comments, and conduct pattern analysis in the conversation, and analyze and compare the results.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
