Highly Intelligent Recommendation Algorithm based on Matrix Filling
AUTHORS
Andrew M. Barthelemy,University of Adelaide, Adelaide, Australia
George Suter,University of Adelaide, Adelaide, Australia
ABSTRACT
In the era of big data, how to solve the problem of information overload is very important. Compared with information systems, search engines, and other technologies, personalized recommendation technology has the advantage of precise positioning. Based on the analysis of the classic collaborative filtering algorithm, this paper proposes a personalized recommendation algorithm for the problem of the decrease in the quality of recommendation results caused by the sparseness of the data. First, use the Slope One algorithm selected by the optimization project to fill the user-score matrix. Then use the filled matrix to form a Top-N recommendation list, and finally filter the recommendation list again according to the user's interest preferences to get the final recommendation result. The experiment verifies that the algorithm in this paper can alleviate the problem of data sparsity through comparison with other collaborative filtering algorithms, and can effectively improve the recommendation quality of the system when the scale of data and its sparsity continue to increase.
KEYWORDS
Collaborative filtering algorithm, Personalized recommendation, Data sparsity, Matrix filling