Collaborative Filtering Recommendation Algorithm based on Improved Similarity

The traditional collaborative filtering recommendation algorithm can't measure well the similarity among users under the condition of data sparsity, which leads to a decrease in the accuracy of the recommendation system. Therefore, we propose a collaborative filtering algorithm to improve similarity. We weight the number of common scoring items on the similarity calculation to improve the shortcomings of the usual Pearson correlation coefficient. Then, we use it to find the user's nearest neighbor set and predict the score of the item to generate a recommendation list. Experiments show that our algorithm can improve the accuracy of system recommendation effectively.