Predicting wet-season rice yield of Gangetic West Bengal through weather based regression model using dummy variable
Keywords:
Wet-season rice, Dummy variables, Multiple regression, Yield forecastingAbstract
Dummy variable multiple regression technique was employed to evaluate the integrated effect of date of transplanting (qualitative variable) and weather factors (quantitative variable) to forecast the wet season rice yield. The regression analysis proposed the methodology for conversion of date of transplanting to the quantitative variable introducing dummy variables. The results revealed that only four parameters, namely, minimum temperature, wind speed, relative humidity and two categories of date of transplanting (dummy variable) can predict the variation of 74% of rice yield. The Pearson’s correlation between minimum temperature and wet season rice yield was found highest (r = 0.198). It was observed that the model performance is quite satisfactory (the accuracy is more than 90 %). It can be concluded that such model using dummy variable can be used for more accurate total yield prediction of wet season rice in the district level.
