Real-World Machine Learning

Appendix. Popular machine-learning algorithms

Name	Type	Use	Linear/nonlinear	Requires normalization
Linear regression	Regression	Model a scalar target with one or more quantitative features. Although regression computes a linear combination, features can be transformed by nonlinear functions if relationships are known or can be guessed. R: www.inside-r.org/r-doc/stats/lm Python: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html#sklearn.linear_model.LinearRegression	Linear	Yes
Logistic regression	Classification	Categorize observations based on quantitative features; predict target class or probabilities of target classes. R: www.statmethods.net/advstats/glm.html Python: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html ...