Study on machine learning Algorithms for optimization in web Page recommendation system

Abstract

Recommendation system is one among the predominant applications of data analytics that predicts and recommends most relevant digital items to the end users. There are many classes of recommendationsystems available such as product, book, movie, news recommendation, etc.Web page recommendation is one among the most challenging recommendation systems under the area of service computing. Today, in the internet world of abundant information, web pages are being recommended using the access hit rate of multiple users. This may not be always suitable and relevant for a particular user s search. Hence intelligent recommendation engines are developed that employ machine learning algorithms to predict the user s interest. This leads to the emergence of web page personalization during recommendation process. Personalization can be achieved by focusing the investigations on user s navigation patterns and access paths from their past navigation history. User profiles are created based on these search history which plays a vital role to improve personalization. Several novel algorithms are being developed to improve the efficiency and accuracy of recommender systems. Existing models for recommendation employ various algorithms that are broadly classified as Collaborative filtering, Content-based approach, Association rule mining, Clustering, Classification, Hybrid systems, Evolutionary and Swarm intelligence based algorithms. Various drawbacks are found in the existing models which include cold start problem, prone to popularity bias, sparsity, time consumption, incapable to handle dynamic web pages and lack of newlinepersonalization. newline newline