PLANT DISEASE FORECASTING MODELS DRIVEN BY ARTIFICIAL INTELLIGENCE

Abstract

In reducing losses which are incurred during cultivation of crops, and also promoting a friendly environment by farming, diseases that affect plants should be detected and anticipated within a very short time.  This research demonstrates how it is possible to predict plant diseases using AI through the training and applying machine learning and deep learning algorithms such as Random Forest, XG-Boost, Support Vector Machines (SVM), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM) networks to various datasets.  The algorithms were trained on the photos of the symptoms of the diseases, weather conditions, and past histories of the disease eruptions. This enabled powerful multi-factorial prediction to be made.  These findings indicate that ensemble techniques such as XGBoost and deep learning approaches such as CNN will always give higher accuracy (>95 95 ), precision, recall, and F1-scores as compared to classical classifiers.  While examining the growth of various crops and types of diseases we found that the relationship between them and the predictions when environmental variables were added to the mix were highly improved.  The examples of hybrid affiliations are bar, line, scatter, pie, and multi-plot kids all of which assisted us to realize better the patterns of illness and better ways models operate.  The case study demonstrates that AI-based forecasting of diseases will assist the farmers in making better decisions, having a more prompt response, reducing the amount of pesticides used, and making crops resistant to diseases.  This outcome will unlock the possibilities of advanced, scalable agricultural models that are capable of coping with the issues generated by climate change and novel plant diseases..