Research On Clustering Recommendation Algorithm Based On Cuckoo Search

Faced with the rapid development of the Internet and the explosive growth of information,it is difficult for people to screen valuable resources,obtain the information they need,and the resulting “information overload” is becoming more and more serious.The recommended system has become an important means of e-commerce that can help users quickly and efficiently find and obtain the information they need from massive data,improve user experience and prevent user loss.As a key part of the recommendation system,the recommendation algorithm has received extensive attention from experts and scholars in many fields.The performance of the algorithm directly affects the accuracy of the recommendation results recommended by the system.Therefore,scholars have proposed a number of superior performance recommendation algorithms to provide users with more accurate and efficient recommendation services.Collaborative filtering algorithm(CF)is the most widely used recommendation algorithm in the recommended system domain,but there are still many problems such as data sparsity,algorithm scalability and cold start.In recent years,many scholars have introduced clustering algorithms in data mining into the recommendation system to mitigate these problems by clustering users or projects.The cluster-based collaborative filtering algorithm can solve the problem of data sparsity and algorithm scalability to a certain extent,and has high recommendation accuracy and superior algorithm performance.In this paper,the clustering-based collaborative filtering recommendation algorithm is systematically studied.According to the performance and its advantages and disadvantages of the algorithm,the optimization process is proposed and improved.This will result in superior and more accurate recommendations.Based on the original cluster-based recommendation algorithm,after analyzing the influence of clustering results on the accuracy of recommendation results,a new bionic group intelligent optimization algorithm,Cuckoo Search(CS),is introduced.The characteristics of global optimization and fast convergence are used to improve the local optimal problem that may be caused if the initial clustering center of the clustering algorithm is not well selected.A K-means & Cuckoo Search optimization recommendation algorithm model is designed based on the cluster recommendation algorithm based on cuckoo search.Then,at the time of clustering,the users who choose to participate in a number of scores are set as the initial clustering center,and the cuckoo search algorithm has a slower convergence rate in the later stage,and the adaptive improvement of the step size mechanism is carried out.Finally,a cluster recommendation algorithm based on cuckoo search is implemented in the Movielens dataset.Compared with several other recommended algorithms,it has better results.The proposed algorithm can alleviate data sparsity and algorithm scalability to a certain extent.The recommendation accuracy is also improved,and it also has better recommendation efficiency.