Forecasting Yield of Major Crops in Assam: A Time Series Approach

Nishan Bhuyan *

Department of Agricultural Statistics, Assam Agricultural University, Jorhat-785013, Assam, India.

Supahi Mahanta

Department of Agricultural Statistics, Assam Agricultural University, Jorhat-785013, Assam, India.

Deepak Sarma

Department of Agricultural Statistics, Assam Agricultural University, Jorhat-785013, Assam, India.

*Author to whom correspondence should be addressed.


Abstract

This research endeavors to predict the yield of five major crops in Assam—Rice, Wheat, Potato, Rapeseed & Mustard, and Arhar employing Autoregressive Integrated Moving Average (ARIMA) models. A time series dataset spanning 50 years from 1973-74 to 2022-23 pertaining to crop yields was collected and scrutinized to identify the most suitable ARIMA models for each crop, grounded in model diagnostics and goodness-of-fit criteria. The models selected—ARIMA(0,1,1) for rice, ARIMA(2,1,0) for wheat, ARIMA(2,2,1) for potato, ARIMA(3,1,1) for rapeseed & mustard, and ARIMA(2,1,2) for arhar—served as the basis for yield projections extending to the year 2030. The results indicate a relatively stable yield trajectory for rice and wheat, whereas a pronounced upward trend is evident in potato yields. Rapeseed & mustard and arhar demonstrate moderate growth patterns. The expanding confidence intervals denote increasing uncertainty over time, underscoring the necessity for ongoing model revisions. This study underscores the efficacy of ARIMA models in agricultural planning and policy formation by supplying dependable forecasts for strategic decision-making. The findings derived can facilitate more efficient resource allocation, food security planning, and climate-resilient agricultural development in Assam.

Keywords: Box-Jenkins methodology, stationary, AIC, BIC, forecasting


How to Cite

Bhuyan, Nishan, Supahi Mahanta, and Deepak Sarma. 2025. “Forecasting Yield of Major Crops in Assam: A Time Series Approach”. Journal of Scientific Research and Reports 31 (6):932-42. https://doi.org/10.9734/jsrr/2025/v31i63187.

Downloads

Download data is not yet available.