Table 9 ML learning, techniques and evaluation metrics

1.

Supervised Learning: Models trained on labelled data to predict outcomes

• Regression: Support Vector Machines, Neural Networks, Ridge Regression, Lasso, Random Forest

• Classification: Decision Trees, Random Forest, Support Vector Machines, Discriminant Analysis, Naïve Bayes, Nearest Neighbour, Neural Networks

Accuracy: Proportion of correct predictions or overall correctness

Sensitivity (Recall): Measures the model’s ability to correctly identify positive cases

Specificity: Indicates the model’s ability to correctly identify negative cases

Positive Predictive Value (Precision): Helps understand the likelihood that a positive prediction is correct

F1 Score: Harmonic mean of precision and recall

ROC-AUC (Receiver Operating Characteristic—Area Under Curve): Assesses the trade-off between sensitivity and specificity. Think of it as an overall ability to distinguish between positive and negative cases

2.

Unsupervised Learning: Models that identify patterns in unlabelled data

• Clustering: K-Means, K-Medoids, Fuzzy C-Means, Hierachical, DBSCAN, Gaussian Mixture, Hidden Markov Model, Neural Network),

• Dimensionality Reduction: PCA, LDA, Isomap, Autoencoder

3.

Semi-Supervised Learning: Combines a small amount of labelled data with a large amount of unlabelled data during training

Both of the above

4.

Reinforcement Learning: Models learn by interacting with an environment to achieve a goal

Markov Decision Process