PRE HARVEST PRICE FORECASTING OF BENGALGRAM PRICE IN TELANGANABY USING ARIMA MODEL

Abstract

Accurate pre-harvest price forecasting is essential for enabling farmers to make informed production and marketing
decisions and for improving the effectiveness of agricultural market intelligence. The present study aimed to develop an appropriate
Autoregressive Integrated Moving Average (ARIMA) model for forecasting Bengalgram prices in Telangana using monthly market
price data collected from April 2002 December 2025. The Box–Jenkins methodology, comprising model identification, parameter
estimation, and diagnostic checking, was employed using SAS software. Several ARIMA models were evaluated based on Akaike
Information Criterion (AIC), Bayesian Information Criterion (BIC), parameter significance, and residual diagnostics. Among the
candidate models, ARIMA (112) was selected as the best-fit model, exhibiting the lowest AIC (4075.07) and BIC (4087.66)
values. The residual diagnostics confirmed that the model residuals were white noise, indicating an adequate model fit. The model
forecasted Bengalgram prices in the range of INR 5,100–5,400 per quintal during the pre-harvest period. Validation of the
forecasts against observed market prices showed a percent deviation of only “0.96%, demonstrating high forecasting accuracy.