Hybrid Machine Learning Model for Solar Radiation Forecasting and Plant Care Optimization

Abstract — Accurate solar radiation forecasting is essential for enhancing agricultural productivity and optimizing plant care. The traditional forecasting models struggle with accuracy due to complex weather patterns. In this study, we predict solar radiation using a hybrid machine learning model integrating Convolutional Neural Network (CNN) and CatBoost to enhance solar radiation prediction and provide plant care recommendations. This model processes weather data including temperature, humidity, wind speed, precipitation, cloud coverage and solar radiation. The CNN captures patterns in sequence, while CatBoost handles structured data and it improves prediction accuracy by improving Root Mean Squared Error (RMSE) compared to single model implantation. This model provides a weighted forecast and recommends plant care actions based on optimal solar radiation levels. Future work aims to integrate IoT data and expand geographical coverage to enhance prediction accuracy.