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