Integrating Machine Learning with Association Rule Mining for Enhanced Crop Disease Classification

One of the biggest issues facing the agriculture industry is crop disease, resulting in large losses in crop yield and quality. Early identification of crop diseases and understanding their triggers are vital to food security and agricultural sustainability. Over the years, various Techniques for data mining and machine learning have been in use to analyze crop health and predict diseases however, they still exhibit certain limitations. This paper proposes a model using Association Rule Mining (ARM) techniques, specifically the Apriori Algorithm and FP-Growth Algorithm, to analyze associations between crop conditions and disease occurrence. This approach first collects structured data from the Paddy Doctor dataset, which includes crop variety, plant age, image ID, and disease labels. Association rule mining is applied using both Apriori and FP-Growth algorithms to identify frequently occurring sets of items and generate association rules that reveal hidden patterns between crop conditions and disease manifestation. These rules guide the feature selection process, isolating the most informative features for disease prediction. The selected features are then used to train classifiers for supervised machine learning models and logistic regression. Experimental evaluation demonstrates that Logistic Regression achieves 92.80%, highlighting The efficiency of the suggested method in early crop disease detection.