Modeling Cassava Yield In Western Kenya: Optimal Scaling Integrated With Principal Component Regression Approach

Abstract

Cassava is a major food crop grown in the tropical and subtropical parts of the world.  In this research work, we sought to develop a model for predicting cassava yield using the PCR model integrated with optimal scaling. Moreover, establishing relationship between the different factors of production, estimate the yield based on the key components adduced to the factors of production in trial data in Western region, Kenya.  Principal component analysis and optimal scaling were used. Pearson correlation prior to principal component analysis indicated significance correlation among the factors of production. A prior to principal component regression, analysis using the variance inflation factor also indicated correlation in key factors of yield forecasting, VIF of 1666.667 (R²=0.999). The coefficients derived from this model were unstable and therefore not reliable for yield prediction .Using the amount of explained variance criterion (70%-80%), we selected the first eight principal components which accounted for almost 70% of total model variance. Eight (8) key components were obtained as key determinants of yield; the most vital component having an eigen value of 2.149 and the least important having an eigen value of 1.005. The post principal component regression model was fitted. The PCR model indicated non-correlation among the eight principal components with the VIF attributed to the overall PCR model being 2.564, (R²=0.610 (Adj R²=0.590). The model offers an efficient alternative to existing models for crop yield prediction when the number of factors to be included in the model is high.

Keywords: PCR, PCA, VIF