Learns from labeled knowledge (enter → right output)
Predict a class
“Is that this electronic mail spam or not?”
Algorithms:
- Logistic Regression
- Determination Tree
- Random Forest
- Assist Vector Machine (SVM)
- Neural Networks
- k-NN (k-Nearest Neighbors)
Predict a quantity
“What’s going to the home value be?”
Algorithms:
- Linear Regression
- Ridge/Lasso Regression
- Determination Timber (for regression)
- XGBoost
- Neural Networks
Learns from unlabeled knowledge (solely enter, no output)
Group related issues collectively
“Group related clients”
Algorithms:
- k-Means
- DBSCAN
- Hierarchical Clustering
- Gaussian Combination Fashions (GMM)
Simplify knowledge by decreasing options
“Visualize knowledge with 1000 options in 2D”
Algorithms:
- PCA (Principal Element Evaluation)
- t-SNE
- UMAP
- Autoencoders (neural community model)
Discover relationships or guidelines in knowledge
“Individuals who purchase X additionally purchase Y”
Algorithms:
Learns by reward and punishment
Be taught the worth of being in a state
Q-Studying (well-known for recreation AI)
Be taught a coverage straight (what motion to take)
Algorithms:
Combines each worth and coverage studying
Utilized in superior robotics and video games
Algorithms:
- A3C
- DDPG
- PPO (Proximal Coverage Optimization)
