Research on Book Recommendation Algorithm Based on Improved ALS Collaborative Filtering
In order to solve the problems existing in traditional collaborative filtering algorithms,i.e.,data sparsity,accuracy and quantification of book preference,and obtain more ideal book recommendation effect,this research is based on the algo-rithm principle of the ALS model.The average borrowing time of the reader's book borrowing and returning records is used to generate the reader's book preference matrix,Pearson similarity is introduced to analyze the reader's book similarity,and the problem of factor information loss caused by the ALS model is improved.The system fills in the non-scored item data,and de-termines the algorithm implementation process to recommend the preferred books for specific readers.This paper verifies the accuracy of the algorithm through experiments.The experimental results show that the RMSE value of the improved ALS algo-rithm reduces 8.2 percentages,and the performance and accuracy of the recommended algorithm are improved.
万方数据
维普
