Practical Implications of ANN

Topics

• Neural-network workflow

  • Data preparation with NumPy, Pandas and Scikit-learn
  • Building a network in Keras or TensorFlow
  • Training with epochs and batch size
  • Evaluation with accuracy, precision and recall
  • Visualising performance metrics with Matplotlib
  • Hyperparameter tuning by grid or random search

Learning objectives

1. Plan, train and evaluate a neural network on prepared data.