A Parsimonious model for prediction of monsoon rainfall in India
A research paper by Gowariker have used multiple and power regression involving 15 independent variables for long range forecasting of monsoon rainfall in India. They have also argued that, when most of the independent variables are 'favourable' almost invariably the monsoon rainfall is normal. In this note we formalize this approach using a parsimonious logistic regression model. The probability of a normal rainfall can be assessed in most cases using only five of the 15 variables.