Font Size: a A A

Based On The Explicit And Implicit Filtering Recommendation Algorithm Research

Posted on:2015-03-09Degree:MasterType:Thesis
Country:ChinaCandidate:Q M DongFull Text:PDF
GTID:2298330452454811Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Now travel recommendation with the development of the information age hasbecome more and more popular and necessary. This paper researches on the tourismrecommendation algorithm is intended to improve the final recommendation resultaccuracy, let the recommended content becomes more meaningful. The article embodiedin the theoretical and practical significance of the recommendation algorithm is accurate,it is more convenient for people to obtain information faster and more accurate. We havedone some work as follows:First of all, for now the travel recommendation algorithm and analysis to understand,the recommendation algorithm are most explicit filtering recommendation form results,and the recommendation algorithm has a known as the recommended way to implicitfiltering algorithm. So we will consider to add an implicit filtering algorithm in theexplicit filtering algorithm in the original, to test the hybrid approach that can get thebetter effect of the recommendation.Secondly, In addition to considering adding implicit filtering algorithm, we againconsider the explicit filtering algorithm can improve the original space. This time we dothe optimization of two aspects for the explicit algorithm, one is to solve the problem ofexplicit filtering algorithm for the original, second point is for the explicit filteringalgorithm with time weight, make the result more reasonable.Finally, for a complete recommendation algorithm must have a cold start problemand the malicious user problems, so we had to solve these two problems in this system, webased on the classical solution are discussed and compared, finally obtains the methodbest fits the systematic treatment of cold start and malicious users.The result of the experiment, implicit filtering algorithm in recommendation resultsor played a very good role, but also time weight added to the explicit algorithm indeed forthe final result is more reasonable. Of course, the cold start problem solving and themalicious user problems to make the system have better security, the final result is moreaccurate.
Keywords/Search Tags:Conditional preferences, Implicit preferences, Recommender system, Contradiction rules, Time weight
PDF Full Text Request
Related items