A Tour of Machine Learning Algorithms - MachineLearningMastery.com

Abstraction: Taxonomy of machine learning algorithms by learning style and similarity

Key points:

• Algorithms grouped by learning style: supervised (labeled data, classification/regression), unsupervised (clustering, dimensionality reduction), and semi-supervised
• Grouped by similarity: regression, instance-based (kNN, SVM), regularization (LASSO, Ridge), decision trees (CART, C4.5), Bayesian, clustering (k-Means, EM), association rules, neural networks, deep learning, dimensionality reduction (PCA, t-SNE, UMAP), ensemble methods
• Deep learning separated from classical neural networks due to massive growth; includes CNN, RNN, LSTM
• Ensemble methods (Random Forest, Boosting, Bagging, Gradient Boosting) are highlighted as very powerful
• Semi-supervised learning is a hot topic especially for image classification with few labels

Connections: Machinelearningmastery · Machine Learning · Supervised Learning · Deep Learning

Source: http://machinelearningmastery.com/a-tour-of-machine-learning-algorithms/