NOISE REDUCTION IN WEB DATA: A LEARNING APPROACH BASED ON DYNAMIC USER INTERESTS

Abstract

Users find it very difficult, if not impossible, 
to narrow their search results to precisely 
meet their evolving interests due to the vast 
amount of material accessible online. Recent 
research has suggested noisy web data 
reduction strategies to handle this issue, 
although most online noise originates from 
data that isn't directly connected to the page's 
content or structure. The article's major point 
is that some of the info shown on a homepage 
is irrelevant to visitors and should be 
removed. Learning the noisy online data 
associated with a user's queries improves their 
quality as a user, resulting in a lower amount 
of profile noise and less loss of critical 
information. user profile data is suggested for

usage on the web. In light of users' ever- 
changing interests, the suggested approach 
takes noise data removal into account. The 
suggested work can only be considered 
legitimate if an experimental design setup has 
been laid out. We compare the findings to the 
techniques that are currently used to reduce 
noise in web data. The experimental findings 
demonstrate that the suggested method takes 
into account the ever-changing user interest 
before noisy data is eliminated. In order to 
improve the quality of a web user profile, the 
suggested work helps to decrease the quantity 
of relevant information that is discarded as 
noise.