Collaborative Filtering Recommendation Algorithm Based on Random Forest Filling

Aiming at the sparsity problem of data set in the field of information recommendation, a collaborative filtering recommendation algorithm based on random forest filling is proposed. First, ID3 algorithm is adopted to construct the decision trees and the random forest is composed of the decision trees. Then, according to the steps of the proposed filling algorithm, the missing values of the data set are predicted using random forest and then be filled. Besides, the mean and the mode filling methods are used for a contrast. Finally, the collaborative filtering recommendation algorithm is tested for the datasets processed by different methods. Experimental results show that compared with the general recommendation algorithm and the recommendation algorithm based on the mean and the mode filling, the proposed recommendation algorithm based on random forest filling significantly improves the F value of the recommendation result and further guarantees the quality of the recommendation result.