An AI-Driven Advanced Decision Support System for Coconut Yield Prediction and Disease Detection

Abstract

The agricultural economy of India's coastal regions, especially Andhra Pradesh, depends heavily on coconut farming. Coconut farming, however, is extremely vulnerable to weather fluctuations, such as erratic rainfall patterns, pest infestations, and plant diseases, which have a substantial impact on crop output and farmers' earnings. Conventional agricultural methods frequently lack access to data-driven insights and predictive analytics, relying instead on manual observation and experience-based decision-making. The AI-Driven Coconut Farming Decision Support System (CFDSS) presented in this study uses predictive modeling and cognitive analytics to assist farmers and agricultural stakeholders. In order to produce region-specific insights at the district and mandal levels, the suggested approach combines several agricultural datasets, such as rainfall records, coconut yield statistics, pest risk information, and geographic boundary data. Rainfall prediction is done using a Long Short-Term Memory (LSTM) deep learning model, and yield estimation and pest risk classification are done using a Random Forest machine learning model. Python-based technologies including Streamlit, Pandas, NumPy, Plotly, and TensorFlow are used in the system's implementation to enable an interactive dashboard with visualization tools and AI-generated farming advisories. Furthermore, coconut plant problems are identified and preventive measures are suggested by an image-based disease detection module. The suggested method improves agricultural decision-making by offering precise predicted insights and practical suggestions, according to experimental evaluation. The promise of intelligent decision support systems in promoting precision agriculture and sustainable coconut farming is highlighted by the integration of artificial intelligence, geospatial analytics, and interactive visualization.