It predicts whether the patient has Heart Disease or not. Here I have used a Dense layer as the input layer with 13 features. It also as 5 hidden layers all with the actuvation function as 'relu' and finally one output layer with 2 neurons and activation function as 'sigmoid'.
Here I have used 'Adam' as Optimizer and loss function as 'sparse_categorical_crossentropy'. It finnaly gives Validation Accuracy as 98.78% and Test Data Accuracy as 99.512%.