Research on Collaborative Filtering Algorithm Based on Spark Platform
Tian Hong-Peng, Chen En-Jie · 2017
Aiming at the problem of time-consuming and low accuracy in dealing with massive data using traditional collaborative filtering algorithm, this paper analyzes the principle of the Item-Based algorithm and the ALS algorithm and parallelizes the two algorithms based on the Spark platform. And fuse the two algorithms. Considering the parallelization performance of recommended algorithm and recommended quality, the comparison experiment were implemented on the open data set MovieLens. Experimental results show that even on a single node, the running time of each algorithm is far less than the traditional single machine implementation. The fusion of two kinds of collaborative filtering algorithms can further improve the recommendation accuracy on the basis of the existing algorithms.