Smart Crop Recommendation and Yield Prediction System Using Machine Learning

Our project, Smart Crop Recommendation and Yield Prediction System Using Machine Learning, is developed to address the key limitations of traditional agricultural decision-making practices. Existing approaches primarily rely on farmer experience, regional advisories, and static guidelines that lack personalization, accuracy, and real-time adaptability. Farmers struggle to determine the most suitable crops for their specific soil and environmental conditions, leading to poor crop selection, resource wastage, and reduced productivity. To overcome these challenges, the proposed system leverages advanced Machine Learning algorithms — specifically Light Gradient Boosting Machine (LightGBM) and Random Forest — to analyze soil parameters (Nitrogen, Phosphorus, Potassium, and pH) and environmental factors (temperature, humidity, and rainfall), providing accurate and data-driven crop recommendations. The system performs multi-class classification on structured agricultural datasets and incorporates feature importance analysis to improve transparency.

本项目为基于机器学习的智能作物推荐与产量预测系统（Smart Crop Recommendation and Yield Prediction System Using Machine Learning），旨在解决传统农业决策实践的核心局限。当前主流农业决策方案主要依托农户种植经验、区域农业指导意见与静态操作规范，普遍缺乏个性化、精准性与实时适配能力。农户往往难以根据自身地块的特定土壤与环境条件筛选出最优适配作物，进而引发作物选择失当、资源浪费与生产效益下滑等问题。为破解上述难题，本系统采用先进的机器学习（Machine Learning）算法——具体为轻量梯度提升机（Light Gradient Boosting Machine，LightGBM）与随机森林（Random Forest）——对土壤参数（氮（Nitrogen）、磷（Phosphorus）、钾（Potassium）与pH）及环境因素（气温、湿度与降雨量）开展分析，从而输出精准且基于数据驱动的作物推荐方案。本系统针对结构化农业数据集完成多分类任务，并引入特征重要性分析以提升决策透明度。