Reasearch Of Mobile Application Reviews Analysis Technology

With the development of mobile Internet, mobile applications are widely used in all aspects of life, making the data in each filed becoming more and more redundant, data analysis techniques also apply to various fields. In this case, this paper first summarizes the current data analysis technologies and its application, and then go through the data mining of mobile application, then this paper develops three algorithms i.e. hot entity discovery, entity-topic generation of reviews and mixed user sentiment computing, by doing this, the paper presents a novel hierarchical model for analyzing mobile APP user reviews. Finally, this paper summarizes the research topics of mobile APP reviews, and proposes a new mobile software detection method in security, based on user reviews.Firstly, according to the current research and data analysis technology widely reported by the media, this paper summarizes the theory of data analysis technology. Including Web data crawling technology, the current popular open sources data analysis platform, data analysis algorithms, as well as data analysis of the popular application scenarios. The summarization of this paper not only provides a more detailed data analysis reference for the relevant research scholars, but also has a guiding role for the future development of data analysis technology.Secondly, based on previous studies, the author develops three algorithms for the analysis of mobile user reviews. Recent literatures have illustrated approaches that can automatically extract informative content from noisy mobile APP reviews, however the key information such as feature requests, bug reports etc., retrieved by these methods are still mixed and what users really care about the APP remains unknown to developers. In this thesis, a novel model SAR consist of three algorithms:Stratify APP Reviews, providing developers information about users’ real reaction toward APPs. SAR stratifies informative reviews into different layers, grouping the reviews based on what users concern, and this thesis develops a method to compute the user general sentiment on each entity.The model performs user-oriented analytics from raw reviews by (ⅰ) first extracting entities from each review, identifying hot entities of the APP that users mostly care about, (ⅱ) then stratifying all the reviews into different layers according to hot entities with a four-layer Bayes probability method, (ⅲ) and finally computing user sentiments on hot entities. This thesis conducts experiments on three genres of APPs i.e.Games, Social, and Media, the result shows that SAR could identify different hot entities refer to the specific categories of APPs. Accordingly,it can stratify relevant reviews into different layers, the sentiment value of each entity can also represent users’ satisfaction well, this thesis also compares the result with human analysis, with the similar accuracy, the SAR model can speed up the overall analysis automatically. The model proposed by this thesis can help developers quickly understand what entities of the APP users mostly care about, and how do they react to these entities.Finally, this thesis analyzes and studies the application of data analysis technology in the mobile field, and proposes a novel mobile application detection approach based on user reviews. Dynamic detection or static analysis technologies are mostly concentrate on the application code data itself, different from the traditional mobile application security detection, this thesis proposes a new mobile software security detection method based on the analysis of user reviews. Extracting the keywords from the user reviews by the Natural Language Processing tool, and building a mobile software security features dictionary library, by keywords matching, the thesis successfully detects all the user reviews that have security risks.