Research On Collaborative Filtering Recommendation Algorithm Based On Association Rule Optimization

The recommendation algorithm recommends the most critical part of the system.The stability and efficiency of the recommended algorithm can determine the accuracy of the recommended results to a large extent.In this paper,we put forward a collaborative filtering algorithm based on the project,and give the prediction value of the user score matrix,and improve the calculation method of cosine similarity.In the traditional collaborative filtering recommendation algorithm,the difficulty of data sparsity is often encountered,and the proposed algorithm successfully solves the problem,thus improving the accuracy of the prediction value of the scoring matrix.Choosing the appropriate confidence and support threshold is the core problem in the association rule recommendation algorithm,so in this paper,we use the particle swarm intelligence optimization algorithm to solve the core problem,in order to improve the accuracy of the recommendation.Through consulting the literature,we know that the filtering recommendation based on user collaboration has advantages and disadvantages,and the Apriori recommendation algorithm based on association rules also has its advantages and disadvantages.However,in this paper,we integrate the two algorithms,standardize and extract the two algorithms by giving different weights to the recommended results of the two algorithms.Advantages and overcome their shortcomings,a parallel hybrid recommendation algorithm is proposed.Finally,using data in the data set,we simulate the new hybrid recommendation algorithm and two basic algorithms,and compare the three algorithms with F1,accuracy,MAE,recall and algorithm efficiency.The results show that the hybrid algorithm proposed in this paper is very good.Sex and practicality.