This project is an attempt to classify the iris dataset using the neural network, created using keras.
The model is designed to have 3 layers with the following characteristics
- Layer 1; input layer 4 units, activation: RELU
- Layer 2; input layer 5 units, activation: RELU
- Layer 3; input layer 3 units, activation: SOFTMAX
Softmax is used to classify the output into the 3 possible classes; Setosha, Virginica and Versicolor at the output layer.
With just a dataset of 150 examples we were able to attain an accuracy of 97.78 %