Project Goals: • Train an Artificial Neural Network (ANN) to classify the handwritten digits in the MNIST dataset. • Showcase the implementation of a feed forward neural network architecture using popular deep learning library Keras. Project Components:
- Data Preprocessing: • Loading and understanding the MNIST dataset. • Data exploration and visualization to gain insights into the dataset. • Data preprocessing steps like normalization to achieve faster convergence of the optimizer.
- Neural Network Architecture: • Defining the number of layers, neurons, and activation function. Activation function used is ReLU.
- Training the Model: • Splitting the dataset into training and testing sets. • Configuring the model training process, including loss functions and optimizers. • Training the model using forward and backward passes (back propagation).
- Model Evaluation: • Evaluating the model's performance on the test dataset. • Measuring accuracy.
- Model Testing: • Using the trained model to make predictions on new, unseen handwritten digits.
- Dependencies: • TensorFlow, Keras, Matplotlib