An ordered weighted operator approach towards web usage mining

Regression analysis is one of the techniques of data mining and is used to generate futuristic model for a given problem. One of the major challenge in the study of Regression analysis is to reduce the Outlier values. But the inherent complex and flexible nature of web data makes it difficult for various regression algorithms to propose an ideal futuristic model. This paper discusses a method that tries to reduce Outlier values by using a variant of Ordered Weighted Operators. First, a regression problem will be formulated and then it will be converted into a Multi-criteria decision making problem. Results, thus obtained will show that the outliers can be reduced to a significant extent by proper use of these operators.