Design And Implementation Of Precision Recommendation Algorithm In Mobile Payment Platform

With the innovation and the development of information technology in nowadays, s mart phones and other mobile devices have been popularized in China. The new equipment brought a variety of business opportunities in mobile Internet of China. For example, location-based social features, map navigation, location-based services business recommendation have been into our lives. Mobile Payment platform is one of them.After the emergence of new markets, Third-party mobile payment platform has ushered in to the fierce competition of various Internet companies. How to increase market share and how to improve user experience have be the most urgent problem needs the mobile payment software to solve. This article aims to propose a personalization recommendation service based on mobile payment platform for enhancing the stickiness of this mobile platform for users and enhancing the activeness of it.Since the reason that the mobile payment platform needs to improve the user experience, this article proposes and realizes a precise recommender system of mobile payment platform. In the design of the algorithm, it uses the personal information data of a mobile payment platform in China and the historical transaction records between these users and merchants as the train data. The system uses collaborative filtering and multiple regression analysis to predict the merchants which the user may transact in the future as the recommendation result. In collaborative filtering algorithm, the system takes the number of transactions as the score of user to merchant depending on the characteristics of the transaction data. The System also changes the similarity formula between two users with the user profile data. It makes the similarity of users more precise. In the design of multiple regression analysis, it also uses the features of the user profile data as the independent variables of regression analysis for improving the accuracy of the predict result.This article designs the algorithm for predict the transaction in the future based on the data of mobile platform and realizes the system. It also designs the experiment to verify the system. The experiment uses the transaction data in first two months and user profile data to predict the transaction in the third month. The experiment verifies the result with three indicators: precision, recall and F1. By analyzing the result, we can conclude that the change Keywords: mobile payment platform; recommendation; collaborative filtering; regression on collaborative filtering algorithm and multiple regression analysis can improve the accuracy of the result. With final fusion algorithm, the system can get the best predict result.