Heart Disease Prediction with an Amharic Chatbot

Predicting the onset of heart or cardiac diseases is a critical challenge in modern medicine, with approximately One Hundred Seventy person death per minute attributed to heart disease. The vast amount of healthcare data underscores the importance of data science in extracting meaningful insights. Multiple data model strategies are explored to evaluate the likelihood of patients developing heart disease and categorize risk levels using chosen classifiers. Data from Menelik Hospitals in Addis Ababa, Ethiopia, is utilized, and multiple performance metrics such as sensitivity, accuracy, AUC, specificity, ROC curve, and F1-score are employed for evaluation. The outcome is a Heart Disease Prediction integrated with an Amharic Chatbot. Additionally, the Random Forest algorithm emerges as the top performer, achieving an impressive 99% accuracy for predicting heart disease. Simultaneously, the Amharic Chatbot, using the Feed Forward Neural Network Model, achieves commendable performance with 92.88% accuracy, facilitating user-friendly communication in Amharic and promoting cultural diversity.