In this era of modern computing and sophisticated technologies, fraudsters are always at the verge of abusing the benefits and features of the web in one or the other way. Web plays a crucial part in today's life. Almost everything has become more of online rather than offline in the world around us. Data is a very vital part of our day-to-day life as we share data with others to communicate and express. These data when uploaded to the web becomes a very important property to secure. Therefore web is very prone to be misused by the fraudsters in order to steal and make use of others data. Hence the web needs to be secured today more than any other time. Data mining is a concept of sniffing through large sets or amounts of data to mine certain patterns to cluster and classify various features of the dataset. Fraudsters follow certain pattern while misusing the web in order to satisfy their greedy needs. Data mining can be applied to extract such vicious patterns by mining through the web data in order to classify certain activities as fraudulent. This paper discuss the use of data mining and the various algorithms in a more hybrid approach in detecting such insecure patterns on web and classify them as fraud by making use of feature analysis and making the web more secure for the people to surf. Copyright © 2018 American Scientific Publishers All rights reserved