The Advantage And Disadvantages Of Web Mining

Abstract—The quantity of accessible information on the web continues to grow rapidly and has exceeded human processing capabilities. The sheer amount of the information increases the complexity for users from discovering desired information. Recommendation systems have then became a valuable resource for users seeking intelligent ways to search through enormous volume of information available to them. Web logs are important information repository which records users activates on search results. The mining of these logs can improve the performance of search engines, since user has a specific goal when searching for information. In this paper, a survey is provided on the different recommendation techniques with their advantage and drawbacks.

To obtain some interesting knowledge, access (browsing) patterns, user behaviors the detailed and in depth study of this data source is essential. Using this better design, structure and content can be created for web site and represented to the user. In contrast to the standard data mining methods need to deal with the heterogeneous, semi structured or unstructured data [1]. Web Mining aims to discover the useful information or knowledge from the web hyperlink structure, page content and usage data. Web mining can be categorized based on the interest what kind of knowledge to be mined from the web data [2]. 1) Web Content Mining: refers to mining or extracting the useful data from the Web pages i.e. text, images and records etc. [3] 2) Web Structure Mining: aims to discovery the knowledge from the hyperlinks which improves the quality of web site.[3] 3) Web Usage Mining: discovers uses access patterns from patterns from the web usage logs which records every click made by the users. Web usage

These are users that have provided similar feedback to a large number of the items that have been consumed by the active user. Items that have been consumed by likeminded users but not by the current user are candidates for recommendation. The assumption made by these systems is that users that have had common interests in the past, will have similar tastes in the future [4]. Demographic Recommendations Categorizes the user based on personal attributes and makes recommendations based on demographic classes, e.g. college students, teenagers, women, men, etc. The advantages is that it does not depend upon any database of users rating and disadvantage of this system is that the knowledge base needs to remain updated to keep an eye on the changing preferences of customer.

The hybrid recommenders have been motivated by the observation that single recommendation techniques have some difficulties which is needed to be overcome. These can be done by the approach of combining two recommended system. The most common form of hybrid recommender combines the content based and collaborative based filtering