Published April 14, 2026
| Version v1
Journal article
Open
A Data-Driven System for Crop Switching Risk and Yield Stability in Climate Change
Authors/Creators
- 1. Department of Computer Science and Engineering ,Government College of Engineering, Bargur– 635104, Krishnagiri, Tamil Nadu, India Anna University, Chennai– 600 025
Description
Intensifying climate volatility poses unprecedented threats to traditional Indian agriculture, disrupting seasonal crop calendars and rendering historical planting strategies unreliable. This paper presents the system - a fully automated, web-accessible, data-driven decision-support platform that quantifies the risk associated with transitioning from one crop to another under climate uncertainty. The system integrates four synergistic analytical layers: (1) a Large-Scale Historical Ingestion Engine processing 50,765 agricultural records spanning 1966–2017 across 20 Indian states; (2) a Feature Signal Synthesis module that engineers the Climate Stress Index (CSI), Crop-Climate Interaction terms, and 5-year rolling Yield Volatility metrics; (3) a Hybrid Neural-Gradient Predictive Engine combining Long Short-Term Memory (LSTM) temporal sequence extraction with Extreme Gradient Boosting (XGBoost) tabular regression, achieving a benchmark coefficient of determination (R²) of 88.16% and RMSE of approximately 281 kg/ha; and (4) a Live Intelligence Layer powered by the Open-Meteo Satellite API and Nominatim village-level geocoding, enabling dynamic, context-aware risk score adjustment. The system generates a calibrated 0–500+ Risk Scoring Engine that classifies every crop-switching scenario into LOW, MEDIUM, or HIGH risk categories based on percentile-derived thresholds. Deployed as an interactive Streamlit analytical dashboard with eight specialised pages, the system delivers localised, village-level pilot transition plans to smallholder farmers. Experimental validation confirms that LSTM temporal embeddings contribute a 6.4% absolute R² improvement over the standalone XGBoost baseline. the system positions itself as a scalable, equitable technological shield for agricultural communities navigating climate-driven crop diversification decisions.
Files
a-data-driven-system-for-crop-switching-risk-and-yield-stability-in-climate-change-IJERTV15IS040791.pdf
Additional details
Related works
- Is identical to
- Journal article: https://www.ijert.org/a-data-driven-system-for-crop-switching-risk-and-yield-stability-in-climate-change (URL)