100 Rice yield forecasting using agro-meteorological variables March 2021 Journal of Agrometeorology 23 (1) : 100-105 (March 2021) The weather variables impact the crop differently throughout the various stages of development. The degree of weather effect on crop yield thus be determined not only by the magnitude of weather variables but also on the variability of weather over crop season. Crop yield forecasting methods incorporating weather information’s provides a better prediction of yield accounting the relative effects of each weather components. Regression analysis is the most frequently used statistical technique for investigating and modelling the relationship between variables. Building a regression model is an iterative process. Usually several analyses are required as improvement in the model structure and flaws in the data are discovered. The use and interpretation of multiple linear regression models depends on the estimates of individual regression coefficients. However, in some situations the problem of multicollinearity exists when there are near linear dependencies between/among the independent variables. The Principle component analysis method has been proposed to address when the problem of multicollinearity. Using Rice yield forecasting using agro-meteorological variables: A multivariate approach GURMEET NAIN 1 , NITIN BHARDWAJ 2 , P.K. MUHAMMED JASLAM 2* , CHANDER SHEKHAR DAGAR 3 and ANURAG 3 1 Division of Agricultural Economics, IARI, Pusa, New Delhi- 110012 2 Department of Mathematics and Statistics, 3 Department of Agricultural Meteorology, CCS HAU, Hisar, Haryana- 125004 *Corresponding author email: jaslam.stat@hau.ac.in The weather variables impact the crop differently throughout the various stages of development. The weather effect on crop yield thus can be determined not only by the magnitude of weather variables but also on the variability of weather over crop season. Crop yield forecasting methods incorporating weather information provide a better prediction of yield accounting the relative effects of each weather component. Regression analysis is the most frequently used statistical technique for investigating and modelling the relationship between variables. Building a multiple regression model is an iterative process. Usually several analyses are required for checking the data quality as well as for improvement in the model structure. The use and interpretation of multiple linear regression models depends on the estimates of individual regression coefficients. However, in some situations the problem of multicollinearity exists when there are near linear dependencies between/among the independent variables. The Principal Component Analysis (PCA) method has been proposed to address the problem of multicollinearity. Using principal component scores (PC) derived from weather variables as predictor variables helps to obtain better estimate the yield. The discriminant analysis is a multivariate technique involving the classification of separate sets of objects (or sets of observations) and assigning of new objects (or observations) to the groups defined previously. Forecasting of crop yield can also be done using discriminant analysis scores based on the weather variables as regressor. Keywords - Multiple linear regression, principal component analysis, discriminant function analysis, pre-harvest forecast, crop yield and weather variables principle component scores derived from weather variables as predictor variable helps to obtain better estimate the yield. The discriminant analysis is a multivariate technique involving the classification of separate sets of objects (or sets of observations) and the assigning of new objects (or observations) to the groups defined previously. Forecasting of crop yield can also be done using discriminant analysis scores based on the weather variables as regressor. Rai and Chandrahas (2000) used the discriminant function analysis of weather variables to develop statistical models for pre-harvest forecasting of rice-yield in Raipur district of Chhattisgarh. Bal et al. (2004) developed agro- meteorological wheat yield forecasting models using multiple regression technique for the Ludhiana district. Regression models were found to perform better by combining technology trend with weather variables (Mallick et al., 2007). Agrawal et al. (2012) have derived prediction models for wheat yield in Kanpur district (U.P.) using discriminant functions study of weekly weather data. To overcome the difficulty of multicollinearity observed among plant biometric