Personalized mobile app recommendation method

Abstract

The invention relates to a personalized mobile application APP recommendation method. The present invention obtains the information of users and APPs from the application market, and preprocesses them, uses the emotion-aspect-region model, inputs the preprocessed documents, and obtains the user's potential preferences for the emotion-aspect-region of the APP respectively, Predict the probability value of a user’s selection of an APP, convert it into an APP index file and a user index file after processing, use the co-related topic model to obtain the User‑App recommendation score matrix, and combine the probability value obtained by the above SAR model with that obtained by the CTM model The recommendation scores are linearly combined to assign weights to achieve the final recommendation value. The invention overcomes the defects of the traditional recommendation method that only considers a single element. The present invention comprehensively considers the aspects, emotions, categories and regions in the comments to discover the potential preferences of the users, which is more in line with the actual needs of the users, explores the user's preference for each attribute of the APP, and better understands the needs of the users and the characteristics of the APP. Cold start problem.

Description

technical field [0001] The invention proposes a mobile application recommendation method, in particular to a personalized mobile application APP recommendation method. Background technique [0002] The development of mobile apps has given users more convenience and facilitated their lives. However, the innumerable number and variety of APPs have also caused some problems for users. The study found that it is quite difficult to find useful and user-preferred APPs only through browsing and simple queries. To some extent, excessive information means lack of information. Therefore, some kind of tool is needed to quickly find what users need and prefer. information to assist decision-making and prevent users from getting lost. As a result, many APP recommendation methods have emerged. [0003] Before the present invention was made, traditional recommendation methods such as collaborative filtering technology (CF) were dedicated to finding users with similar interests to the sp...

AI Technical Summary

Problems solved by technology

However, as personalized business applications extend to all aspects of user life information flow, personalized recommendation technology is also developing rapidly. Early technologies such as collaborative filtering can no longer meet the requirements of new environments, such as user and commodity In the case of more and more cases, the performance of the system will be lower and lower, or when the user's evaluation of the product is very sparse, the similarity between users based on the user's evaluation may be inaccurate or even cause the product not to be recommended

Images