The Research On Personalized Recommendation Algorithm Based On Comprehensive Evaluation

With the rapid development and popularization of the e-commerce and explosion of Internet information, problem of information overload is more and more serious. It’s more difficult that the users who find the goods that they need in the vast amount of products information. Personalized recommend system can effectively resolve the problem with interacting with the users and recommend goods basing on their preference and interest, which improves satisfaction and increases the sales of commodities. However, the problems of accuracy, effectiveness and universality seriously restrict the development of recommender system.Serviceable recommendation algorithm not only has the accuracy, but also considers effectiveness of the recommendation and the technology to realize the feasibility. Tradition recommendation algorithm more or less has some disadvantages. At the same time, the new method and the improved technology promote the development of recommendation algorithm. Due to the complexity of the recommendation algorithm in reality, algorithm limits the operation of the network platform. It’s more difficult for small business networking platform to build their recommender system.The main research work of this paper is as follows:1.We review the development of the current recommender system and recommender algorithm, focusing on concluding the basic principles of significant recommendation algorithm and existing advantages and disadvantages, so as to lay foundation for the following.2.In order to update the result of recommender timely, this thesis presents a recommender model based on comprehensive evaluation.This model summarizes the connotation of certain characteristic of an object and divides attribute indicators to accumulate comprehensive weighted summation score. The model uses the method of Top-N to recommend. According to this model, this thesis introduces the process and the mechanism of implementation.3.In this thesis, taking the tourism industry as an example, we build comprehensive evaluation recommendation system. Compared to the collaborative filtering, we design to test the algorithm results.