Classifying Dairy Farmers’ Knowledge, Attitudes, and Practices towards Aflatoxin Using Bootstrap Oversampling and Machine Learning

Aflatoxin is the main naturally occurring mycotoxin that harms animal and human health due to its carcinogenic nature. It is one of the main elements influencing the safety and quality of regional crops and food commodities pre- and post-harvesting. This study aims to analyze the dairy farmer’s data and predict the farmers’ knowledge, attitudes, and practices (KAP) towards aflatoxin using machine learning techniques to enhance intervention strategies. We collected 140 records through questionnaires, divided into training and testing sets with 80% and 20%, respectively. Then we conducted a random over-sampling to balance the data and bootstrap techniques to improve the model's performance. The model achieved an accuracy of 86%, 96%, and 89% for knowledge, attitude, and practice respectively. This indicates that using machine learning to classify dairy farmers' KAP levels is promising. Across all KAP aspects, farm type was identified as an important feature for predicting KAP.