Research On Application Of Association Rule In Mobile Commerce Recommendation System

Mobile commerce is implemented on the wireless platform of e-commerce. Due to itsfast and convenient way of payment and the advantage of providing services anytime andanywhere, mobile commerce is makes the fast progress in the last few years, and as thepopularity of smart phones, mobile commerce has great development potential. Comparedwith the traditional commerce in the PC platform, mobile platforms on the screen can showthe commodity information are relatively limited, how to make the users to quickly find theproducts they are interested in and needed, rather than getting lost in a large number ofcommodity information,has become a mobile commerce development problems which needto be solved. Recommendation system is an effective way to solve the problem of overloadinformation, through the analysis of the interest preference, the personalized needs of users, ituse association rules or collaborative filtering recommendation technology to recommend thepersonalized information to users. however, the existing mobile commerce recommendationsystem still exist some problems, such as sparse, extensible, model fitting problems, leadingto recommend low efficiency, the recommend quality is not high, can’t meeting thepersonalized needs of users, etc. Therefore, the research of mobile commerce recommendersystems and recommendation technology has a large practical value.According to the above problem, this thesis studied the recommendation system andassociation rules theory, analyzed the characteristics of the mobile commerce recommendersystem, working process and organization structure, designed the model of mobile commercerecommendation system based on association rules. In this model, the recommended processis divided into offline and online recommendation. Offline processing is divided into datapreprocessing and association rule mining two sub modules. Data pretreatment use thedatabase triggers and stored procedures to implement the data cleaning and format conversion.Association Rules Mining use FP-growth algorithm for mining frequent patterns, generateand import association rules library, reducing the time of scanning database, and improve themining efficiency. Online recommendation module according to the collection of userinformation and generate the rule base to produce accurate, real-time personalizedrecommendation results. The model recommended in recommended efficiency, quality, hascarried on the beneficial research on real-time and reliability, recommend model in theapplication of mobile commerce system can well referring users to comply with its interestpreferences and demand of goods, so as to improve sales and customer loyalty.