Modeling of Sugar Production in India based on Principal Component

Abstract

Sugar as a luxurious product now becomes an essential one. Sugarcane is the leading cash crop in India used for sugar extraction. Sugarcane was cultivated on 4 million hectares of land shows increasing trend of production over last 10 years, in India. Sugar production scenario for Pre-WTO and Post-WTO period in India, is studied. The effect of eight predictors on sugar production in Pre-WTO (1985-86 to 1994-95) and Post-WTO period (1995-96 to 2008-09) is analyzed. Matrix plot exhibits the scatterness in predictor and response variables. Principal component analysis (PCA) is used to reduce the dimensionality of the data. Multiple regression models, for sugar production on combined score of selected predictors, are obtained. The significance of auto-correlation in residuals of sugar production in India using multiple regressions is tested by Durbin-Watson test. The analysis reveals that in Pre-WTO, Post-WTO and overall period sugar production in India increases as a combined effect of four selected predictors. Durbin-Watson test indicates that autocorrelation in residuals is insignificant in these periods. The calculations of uncrushed sugarcane in Pre- WTO period is of the order of 50 percent which reduces to 40 percent in Post-WTO period. Keywords - Autocorrelation, eigen values, matrix plot, principal component, residual.