Forecasting of Sweet Potato (Ipomoea batatas L.) Prices in India
DOI:
https://doi.org/10.48165/IJEE.2022.58203Keywords:
Market intelligence, Price forecasting, Exponential smoothing model, Seasonal autoregressive integrated moving average, Time delay neural networkAbstract
Due to the semi-perishable nature of sweet potato the price fluctuation occur based on demand and supply. Hence, it becomes necessary to precisely forecast market price of sweet potato. Price forecasting of sweet potato was carried out for six selected states in India using time series monthly market price, collected from AGMARKNET price portal from January 2010 to December 2021. Exponential Smoothing Models (ESM), Seasonal Autoregressive Integrated Moving Average (SARIMA) model and Time Delay Neural Network (TDNN) model were used for forecasting of sweet potato price. It was observed that among the forecasting models, the TDNN model predicted accurate future prices of sweet potato based on the lowest Mean Absolute Percentage Error (MAPE), Mean Absolute
Error (MAE) and Root Mean Square Error (RMSE) than SARIMA and ESM. The forecast indicated that the average market price of sweet potato in selected states of India viz., Kerala, Odisha, Gujarat, Karnataka, Maharashtra and Telangana, would be in the range of Rs. 684 to Rs. 2757 per quintal during January 2022 to December 2022. The forecasted price of sweet potato would provide valuable information to the sweet potato growers, government institutions and other stakeholders in the sweet potato value chain to take appropriate decisions on production, marketing and consumption of sweet potato.