Artificial Flora Optimization: A Novel Feature Selection Approach for Heart Disease Prediction

Finding the best solutions for many real-world optimization problems in engineering, science, business, and economics can be quite tricky. These problems are often complex and computationally demanding, making them difficult to solve using traditional methods. In recent years, however, algorithms inspired by nature have emerged as powerful tools for tackling these tough challenges. These nature-inspired approaches, known as meta heuristics, have proven effective in finding optimal solutions to complex problems, offering a promising alternative to conventional methods. Heart disease, especially Cardio Vascular Disease, is a major killer, responsible for over half of cardiovascular deaths worldwide. Predicting the disease in its early stage is crucial for taking preventive measures or seeking treatment. To address this, a new approach called Artificial Flora Optimization has been proposed. Artificial Flora Optimization (AFO) algorithm is used to select the most informative features for identifying heart disease. AFO’s effectiveness is tested by combining it with four different machine learning methods: Logistic Regression, Naive Bayes, XGBoost, and Artificial Neural Networks (ANN). Using data from the Cleveland Heart Disease dataset, AFO coupled with ANN achieved the best performance compared to the other combinations. This suggests that AFO could be a valuable tool for identifying relevant features, especially when used with ANNs, potentially aiding in more accurate disease diagnosis.