Crop Yield Predication Using Extreme Machine Learning for Sustainable Agriculture

Abstract: Sustainable agriculture plays a vital role in ensuring global food security, environmental protection, and economic stability. Accurate crop yield prediction enables farmers and policymakers to make informed decisions regarding crop planning and resource management. This paper proposes a machine learning approach for crop yield prediction using the Extreme Learning Machine (ELM), a fast-neural network model with strong generalization capability. The dataset includes soil nutrients such as Nitrogen, Phosphorous, and Potassium along with environmental parameters like temperature, humidity, and soil type. Categorical variables are encoded and missing values are handled using median replacement. The ELM model uses a Single Layer Feedforward Network with randomly generated input weights and output weights calculated using the Moore-Penrose pseudoinverse. Performance is evaluated using RMSE and R² Score. A Streamlit-based interface is developed to allow users to input parameters and receive instant crop yield predictions.

Keywords: Crop Yield Prediction, Extreme Learning Machine (ELM), Machine Learning, Sustainable Agriculture, Soil Parameters, Climatic Factors, Precision Agriculture, Agricultural Technology, Streamlit, Real-time Prediction