Research And Design Of Intelligent Online Shopping Recommendation System Based On Hadoop

With rapid development of Internet technology, recommender system has become the most important filter tool to solve information overload problem, and help users to find the useful content to them from massive data quickly and high effectively. But in practical applications, types of items and number of users may be very large, the generation and accumulation speed of log information in system continue to grow rapidly, and traditional recommendation algorithm used to run on single machine, and limited by its performance, it’s not far enough to meet the needs of recommendation computation on mass data.In order to solve the problem which traditional recommendation systems can’t quickly process large data.In this paper study, running mechanism of Hadoop and programming framework Map-Reduce, combine in-depth analysis of collaborative filtering recommendation algorithm, designed and implemented a system applied to online shopping recommendation. The main work has1. Improve Item-Based collaborative filtering recommendation parallel algorithm.Through designing efficient parallel algorithm, compute-extensive procedures are distributed to different processing nodes in Hadoop platform. Taking advantage of parallel computing. We accelerate the response of recommendation. The experimental results show that our proposed algorithm is more efficient and scalable than current parallel algorithms.2. Depth analysis of the project-based collaborative filtering algorithm which implemented by Mahout, studying its demerits and coming up with method to improve it, and at last, conducting simulation experiment to prove the feasibility of improvement.3. Combined with improved project-based collaborative filtering algorithms,through user requirements analysis, system framework design, process design and other sectors, applied to achieve on an intelligent recommendation system of online shopping on Hadoop platform.