Abstract
The data mining methods tell that not all the useful information present on the web is with noise
and user‟s interest. The crucial problem is the level of noise that has been occurred in the useful data is turned
out to be the hindrance for placing the same with dynamic interests. The proposed Noise Web Data Learning
(NWDL) and Suffix Tree Clustering (STC) work take any form of useful information as noisy data but this is
apart from the main web page and suggests in eliminating the noise records which specially focuses on casting
off the noise on the subject of the content material and layout of web statistics. Because the noisy data not only
minimize the level, but it also degrades the performance of useful facts that are assigned to the clients based on
user requests. The current research aims to enhance the greatness of the user profile. The method called Noise
Web Data Learning (NWDL) tool/algorithm is proposed to attain the information regarding the noisy data.
Search engines like Google is an important medium for users to obtain useful data. The request is processed in
the lengthy form which is said to be snippets. Hence the processed information can be organized by a file
clustering concept. The proposed method can be installed on the search engines. The research methodology
offers the need for a file clustering technique and the first approach is evaluated based on this domain. The most
important thing is the clusters are created depending on the snippet's size which is quickly returned by the search
engine. Since the clusters are created based on the snippets, the entire information becomes useful by creating it.
Hence these requirements and conditions can be executed in a linear as well as incremental time is known as
Suffix Tree Clustering (STC). The clusters are generated for the shared set of documents.

Key words: Web Mining, Noisy Data, NWDL, Clustering