5.1 Some common statistical models

There are many types of statistical models. Variables may be nominal (categorical) or interval/ratio data. You may be interested in predicting a continuous interval/ratio variable from other continuous variables, or predicting the probability of an occurrence (e.g. of a species), or maybe the count of something (also maybe a species). You may be needing to classify your phenomena based on continuous variables. Here are some examples:

• lm(y ~ x) linear regression model with one explanatory variable
• lm(y ~ x1 + x2 + x3) multiple regression, a linear model with multiple explanatory variables
• glm(y ~ x, family = poisson) generalized linear model, poisson distribution; see ?family to see those supported, including binomial, gaussian, poisson, etc.
• glm(y ~ x + y, family = binomial) glm for logistic regression
• aov(y ~ x) analysis of variance (same as lm() except in the summary)
• gam(y ~ x) generalized additive models
• tree(y ~ x) or rpart(y ~ x) regression/classification trees

here!!