This assignment aims to implement an artificial neural network (ANN) model using TensorFlow or Keras and evaluate its performance on a chosen dataset. Students are encouraged to explore different architectures activation functions, and optimization algorithms.
Objective:
Implement an ANN model using TensorFlow or Keras.
Evaluate the model's performance on a chosen classification or regression dataset.
Students are encouraged to experiment with various architectures, activation functions, and optimization algorithms.
Tasks:
Dataset Selection:Â
Students need to choose a publicly available dataset suited for classification or regression tasks. Common options include MNIST (handwritten digits), Iris flower classification, or Boston housing price prediction.
Data Preprocessing:
Load the chosen dataset.
Perform necessary cleaning and preparation steps like normalization or standardization of features.
Split the data into training, validation, and testing sets.
Model Building:
Import essential libraries (TensorFlow, Keras, etc.).
Define the model architecture (number of layers, neurons per layer, activation functions, chosen optimizer).
Compile the model by specifying the loss function and optimizer.
Model Training:
Train the model on the training data for a set number of epochs and with a specific batch size.
Monitor the training process by observing the loss and accuracy (or other relevant metrics) on the validation set to avoid overfitting.
Model Evaluation:
Evaluate the trained model on the unseen testing set to assess its generalizability.
Calculate performance metrics like accuracy, precision, recall, F1-score (for classification) or mean squared error, R-squared (for regression).
Report and Discussion:
Summarize the chosen dataset, preprocessing steps, model architecture, training process, and evaluation results.
Analyze the model's performance, including any challenges encountered and potential areas for improvement.
Consider exploring different hyperparameter settings, activation functions, or architectures to potentially enhance performance.