Key Aspects of ML Algorithms
| Algorithm | Task | Supervision? | Model | Typical cost function | Computational burden | Assumptions/comments |
| Naïve Bayes classification | Classification | Supervised | Several (e.g., Gaussian) | Probabilistic | Low | Relies on naïve probability distribution |
| Linear regression | Regression | Supervised | Hyperplane | MSE | Low | |
| Support vector machines | Classification or regression | Supervised | Hyperplane | Classification rate, MSE | Moderate | Handles complex problems |
| Random forest | Classification or regression | Supervised | Tree | Classification rate, MSE | Low–moderate | Is tolerant to overfitting |
| ANN | Classification or regression | Supervised (typical); unsupervised/reinforcement learning (less common) | Neurons connected in layers | Classification rate, MSE | High | Is used for complex problems; may be convolutional or deep |
| k-means clustering | Clustering | Unsupervised | Cluster centroid | Distance to cluster center | Moderate (depends on problem) | Identifies centroids and assigns data to nearest centroid |
| Hierarchical clustering | Clustering | Unsupervised | Dendrogram | Distance between data points | Moderate (depends on problem) | Clusters data by identifying data-points that are similar |
| Principal-component analysis | Dimensionality reduction | Unsupervised | Principal components | Moderate (depends on problem) |
MSE = mean square error.