M M Kodabagi, Savita S Hanji and Sanjay V Hanji, Basaveshwar Engineering College, India

Segmentation is one of the very important strategic tools used by the marketer. Segmentation strategy is based on the concept that no firm can satisfy all needs of one customer or one need of all the customers. The customers are too numerous and diverse in their buying requirements, hence the marketers or companies cannot cater to the requirements of all customers that too in a broad market such as two-wheelers. Cluster analysis is a class of techniques used to identify the group of customers with similar behaviors given a large database of customer data containing their properties and past buying records. Clustering is one of the unsupervised learning method in which a set of data points are separated into uniform groups. The k-means is one of the most widely used clustering techniques used for various applications. The main drawback of original k-means clustering algorithm is dead centers. Dead centers are centers that have no associated data points. The original k-means clustering algorithm with Euclidian distance treats all features equally and does not accurately reflect the similarity among data points. In this paper, an attempt has been made to apply enhanced clustering algorithm which uses similarity measure for clustering (segmentation) of two-wheeler market data. The enhanced clustering algorithm works in two phases; Seed Point Selection and Clustering. The method adapts new strategy to cluster data points more efficiently and accurately, and also avoids dead centers. The enhanced clustering algorithm is found to be efficient in meaningful segmentation of two-wheeler market data. The results of market segmentation are discussed.

Enhanced Clustering, Market Segmentation, Two-wheelers, Similarity Measures, Seed Point Selection.