Design And Implementation Of Online Shopping Mall System Based On Android TV

In May 2015,the State Council has firmly pointed out in "Opinions on Developing E-Commerce and Accelerating the Cultivation of New Economic Drivers”that E-commerce has been the driving force for the development of China's economy,which not only creates new consumption demand but also generates new forms of industries.The statement of our country to the E-commerce is of no doubt right but information overload has turned out to be an urgent problem.The ordinary mall system mainly provides two kinds of means to help users find items of interest from a large number of commodities:sorting according to the classification and search engine,which fails to users' satisfaction and business interests into account.The mall system recommended in this thesis is able to display products to people who are interested in them so as to achieve win-win situation between users and business.Compared to mobile phones,tablets,and other Android devices,TVs have a larger screen which is easier for people to share information.Therefore,the realization of the mall system on the platform of Android TV combined with personalized recommended algorithm is of practical significance.The thesis is based on the mall system and combines recommended algorithms through which functions such as product display functions,personalized recommendations,and popular product recommendation functions are implemented.The main tasks include five aspects:(1)detailedly analyze on the functional demands of the mall system in order to generate overall function and based on Leanback,and appropriate interactive interface for TV is realized according to the characteristics of touch free,large screen and remote control of TV;(2)establish the overall development framework,and grade the level and function modules.Among them,the server side adopts Spring MVC and Hibernate to build the business logic processing framework and data storage solution.The client mainly uses the four components of Android to achieve basic functions;(3)the collaborative filtering algorithm based on Mahout platform is used to implement personalized product recommendation.While collecting user behavior logs,the data is prevented from being cheated.Meanwhile,the algorithm is selected as the core algorithm of the recommendation engine through comparing the accuracy,recall rate and average absolute difference;(4)for the problem of data sparse for collaborative filtering algorithm,this paper adopts the hot product recommendation processing scheme;(5)test the function of the implemented Android TV online shopping mall system,the result of which shows that the system functions properly and the interactive interface operates smoothly which achieves the expected goals.