Nagaraja, G. NSHRUTHI, J2017-08-072017-08-072015-07-14TH-11145http://krishikosh.egranth.ac.in/handle/1/5810027647Agricultural price forecasts are an integral component of trade and policy analysis. As the prices of agricultural commodities directly influence the real income of consumers and it also affects the consumers’ access to food. For the purpose of this study two such agricultural commodities are selected i.e. onion and tomato because of huge demand and price fluctuation. The present study was conducted using ten years monthly prices and arrivals of onion and tomato at Bengaluru and Kolar APMC market, respectively. The data for the above period was collected from Kirshi Marata Vahini website and the selection of markets for the study was done on the basis of maximum quantity of arrivals for the markets. The forecasting models used for onion crop are Auto Regressive Integrated Moving Average (ARIMA), Seasonal Auto Regressive Integrated Moving Average (SARIMA) and Generalized Auto Regressive Conditional Heteroscedastic (GARCH). The forecasting models used for tomato crop are ARIMA, Auto Regressive Integrated Moving Average exogenous variable (ARIMAX) and GARCH. These models were used to forecast prices for next 12 months. In case of onion price, SARIMA model was found to be the best forecasting model by giving a Minimum Absolute percentage error (MAPE) value 21.24 per cent. In tomato crop, ARIMAX model was found to be best forecasting model with a minimum MAPE value of 59.06 per cent. The identification of the best forecasting model would help the producers, consumers as well as suppliers in taking appropriate decisions.ennullThesis