Pipeline to predict heart disease by tuning classical classifiers with a Genetic Algorithm (via NIAPY & PSO) and then refining deep-learning model architectures using AutoKeras neural-architecture search.
The project loads the “heart.csv” dataset into pandas, performs standard preprocessing (scaling, train–test split) , then defines an optimization task over a RandomForestClassifier’s hyperparameters using a Genetic Algorithm from NIAPY and PSO from PySwarm
In parallel, it employs AutoKeras to search for an optimal neural-network architecture (built on TensorFlow) and compares both approaches via accuracy, confusion matrices, and classification reports.