Abstract

In market basket analysis user ranking of products is very important in addition to the values of attributes of
objects. Similarity based comparison between objects play a very important role in many business operations
such as ranking of objects with respect to the preferences of customers, finding a set of top-k objects in ranking
order, and finding a set of k-nearest neighbor objects in ranking order and so on present study proposes a new
similarity finding measure between objects. This measure computes weighted sums of values of attributes and
priority values of respective values of attributes. These weighted sums are computed using a linear function
formula. Finally a new clustering technique is proposed for clustering market basket analysis products using
newly proposed similarity search measure between two objects new clustering technique based on new
similarity finding measure is very useful in many real time applications and in many very large database
operations including query execution.