Research On Recommendation System In E-bussiness Based On The Expert Decision-making Algorithm

In recently years, with the fast development of computer network, Thetransactions on E-commerce platform has been gradually involved into people’s dailylife, the emergence of electronic commerce greatly changed the traditional businessbehavior. Because of the information processing technology and the ever-changingcustomers’ demand,the E-commerce recommendation system on the basis of the firstsingle manual recommended way to gradually integrate into the information retrieval,data mining, artificial intelligence and other means of modern technology.So in thispaper,we propose the study of modern E-commerce recommendation. In the widelyused E-commerce recommendation system,it get the recommended choice combinedwith customer demand by analyzing the real-time browsing and shopping history ofcustomers. E-commerce recommendation system, in accordance with their respectivefields of application to use the appropriate recommendation algorithm, however, thereare two shortcomings exist in the mainly used recommendation algorithm.Firstly,when it comes to the commodities that users do not buy frequently, therecommendation system can not get enough customer information, includingcustomer purchase history, customer assessment, customer interest in preferences, etc.In this way, it is unsufficient to support the implementation of the traditionalrecommendation algorithms; Secondly, for regularly renewed high-techproducts,there is very limited information on the network.With the lack of sufficientprofessional knowledge, customers when facing a large number of high technologicalcontent of goods, it is difficult to give accurate description of the demand, whichbrings the difficulty for customers to make right choice.Directing at the two shortcomings of modern E-commerce recommendation system, this paper proposesthe expert decision-making algorithm used in theE-commerce recommendation system, creating a model based on Multi-Agent System.We can solve the problem effectively by using this model This article studies thetheoretical knowledge of the agent system, including the FIPA standards, thecommunication mechanism between agents and multi-agent system model design.Then introduces three major recommended strategy for E-commerce recommendationsystem widely used in recent, including analysis of their respective basic concepts,algorithms process and their respective advantages and disadvantages and fields ofapplication. Next, the paper proposes an expert decision-making algorithm, and givesa more detailed analysis on its basic principles and algorithms process. Finally, thispaper presents the design architecture of the system and the partial results.Overall, through the deeply study of the modern e-commerce recommendationsystem, we proposed the recommending model of expert decision-making algorithmapplied to a special area to direct at the vacancy of its practical application. Theresearch has the following main points:1. A completed the studies on the current main recommendation algorithm.Analysis the widely used three major recommending strategy: recommendationbased on association rules,recommendation based on content and recommendationbased on collaborative filtering. Done a deeply study of their theoretical basis, thealgorithm processes and their advantages and disadvantages, paving the way for therecommended model of expert decision-making algorithm we are going to introducenext.2. Completed the design and implementation of the recommending model of expertdecision-making algorithm.Detailed analysis the design process of expert decision-recommendationalgorithm model. Completed the design of the experts-knowledge analysis moduleand experts-decision-making module, as well as the initial realization of thealgorithm.3. Completed the design of a Multi-Agent system based on the JADE platform.Expert decision-making algorithm combined to the Multi-Agent Systems, initiallycompleted a multi-Agent recommendation system specifically based on the expertdecision-making algorithm on the JADE platform.The innovation of this paper is: proposed an expert decision-making algorithmmodel based on Multi-Agent system, which can be used in the special field ofe-commerce recommending. Through the research on the application of e-commerce recommendation system,we have proposed and designed a specific application recommendation: algorithm-expert decision-recommendation algorithm. The algorithm is more suitable for therecommendation of goods with newly technology and the user does not buyfrequently.