Web Usage Recommendation System using KNN Classification

Abstract� In today’s world the large problem of many online websites is the presentation of multiple choices to the user at time. So, its very difficult task for user to find desired product or information on the site. Web mining and recommendation system is based on the current users brhaviour and his/her search pattern on the newly developed Really Simple Syndication (RSS) reader website, which provide relevant information to users according to their needs and interest. The k-NN classifier is very simple to understand and implement than other machine learning algorithms. The k-NN classification method is useful in online and real time to identify the users serach pattern, matching it to a particular user group and recommend a smart option that meet the need of specific user at particular time. For this purpose current users previous logs are extracted, cleansed , formatted and grouped into meaningful session to develop data mart.