A Hybrid Regression Model for Cashew Nuts Price Prediction

This paper proposes a hybrid regression model based on the regression tree and multiple linear regression model for improving prediction accuracy and to overcome one of the main disadvantages of the Regression Tree. The performance of the proposed model is compared with regression tree, K-nearest neighbor regression, multiple linear regression, and support vector regression through a Monte-Carlo simulation study. The simulation result indicates that the hybrid model outperforms all other regression models irrespective of sample size when the observations are from a normal distribution and uniform distribution. As an application, the proposed hybrid model is used to solve a problem faced by cashew nuts farmers and buyers to decide the most appropriate prices for the cashew nuts. The results from the hybrid model can be used as a guide by the farmers for fetching better prices in the market and by buyers for getting a lot of ascertained quality.