Machine Learning

You can find Machine Learning exercises that are a continuation of the INSA Toulouse course.

During these exercises, you will find exercises that focus on different topics related to Machine Learning.

We will cover several approaches:

• Supervised approaches:

  • Perceptron (linear regression & gradient descent)
  • Neural networks (Multi-layer perceptron (MLP))
  • Decision trees

• Unsupervised approaches:

  • Clustering methods
    • K-means
    • K-medoids
    • Agglomerative clustering
    • DBSCAN
    • HDBSCAN

In each approach, we will discuss and highlight their characteristics/properties, such as the notion of:

• Similarity/distance: Euclidean, Manhattan, Hamming...
• Evaluation of approaches: MSE, MAE, RMSE, Silhouette score...
• Loss function
• Regression (prediction)/classification
• Cross-validation
• Overfitting
• Activation functions: logistic/sigmoid, Rectified Linear Unit (ReLU), Softplus, tanh...
• Backpropagation
• Pooling