Reduction in Dimensions and Clustering Using Risk and Return Model

Abstract

We introduce a new approach of reducing dimensions and clustering of a database, inspired from a computational model used to evaluate economical parameters. This computational model is based on two well known methods for the valuation of assets, namely the dividend valuation model (DVM) and the capital asset pricing model (CAPM). The model we introduce is called the Risk and return model (RRM), and the technique of dimensions reduction is based on calculating the highest risk or in other words the lowest return associated with each attribute/column in the database. The attributes with the highest risk or lowest return grades are reduced. We have applied a model similar to DVM to cluster the dimensionally reduced data.