Linear Classification in R

 In this post you will discover recipes for 3 linear classification algorithms in R.

All recipes in this post use the iris flowers dataset provided with R in the datasets package. The dataset describes the measurements if iris flowers and requires classification of each observation to one of three flower species.

Let’s get started.

Logistic Regression

Logistic Regression is a classification method that models the probability of an observation belonging to one of two classes. As such, normally logistic regression is demonstrated with binary classification problem (2 classes). Logistic Regression can also be used on problems with more than two classes (multinomial), as in this case.

This recipe demonstrates multinomial logistic regression method on the iris dataset.

# load the package library(VGAM) # load data data(iris) # fit model fit <- vglm(Species~., family=multinomial, data=iris) # summarize the fit summary(fit) # make predictions probabilities <- predict(fit, iris[,1:4], type="response") predictions <- apply(probabilities, 1, which.max) predictions[which(predictions=="1")] <- levels(iris$Species)[1] predictions[which(predictions=="2")] <- levels(iris$Species)[2] predictions[which(predictions=="3")] <- levels(iris$Species)[3] # summarize accuracy table(predictions, iris$Species)

Learn more about the vglm function in the VGAM package.

Linear Discriminant Analysis

LDA is a classification method that finds a linear combination of data attributes that best separate the data into classes.

This recipes demonstrates the LDA method on the iris dataset.

# load the package library(MASS) data(iris) # fit model fit <- lda(Species~., data=iris) # summarize the fit summary(fit) # make predictions predictions <- predict(fit, iris[,1:4])$class # summarize accuracy table(predictions, iris$Species)

Learn more about the lda function the MASS package.

Partial Least Squares Discriminant Analysis

Partial Least Squares Discriminate Analysis is the application of LDA on a dimension-reducing projection of the input data (partial least squares).

This recipe demonstrates the PLSDA method on the iris dataset.

# load the package library(caret) data(iris) x <- iris[,1:4] y <- iris[,5] # fit model fit <- plsda(x, y, probMethod="Bayes") # summarize the fit summary(fit) # make predictions predictions <- predict(fit, iris[,1:4]) # summarize accuracy table(predictions, iris$Species)

Learn more about the plsda function in the caret package.

Summary

In this post, you discovered 3 recipes for linear classification that you can copy and paste into your own problem.