A Novel Dimension Reduction Technique based on Correlation Coefficient

In this paper, a novel simple dimension reduction technique for classification is proposed based on correlation coefficient. Existing dimension reduction techniques like LDA is known for capturing the most discriminant features of the data in the projected space while PCA is known for preservin g the most descriptive ones after projection. Our novel technique integrates correlation coefficient and method of elimination for feature selection to reduce the dimensionality of input space. Our approaches are novel because our method finds alternatives to LDA and PCA in a 2D parameter space. Extensive experiments are conducted on various datasets.