Homework 1

Install R on your system (either unix/linux or windows). If you decide to use machines in 217, then R is already installed.

Read part of "an introduction to R" (available from R homepage) and realize different types of objects (e.g., vector, factor, list, data frame etc.).

Install the package "random forest" (available from CRAN). For this homework, we would like to use this method for classifying the data set "iris." "iris" is available from R by default, so you can use

> data(iris)
to get it. You then randomly separate the iris data to 2/3 for training and 1/3 for testing.

Use random forest to do training and testing, and write a report (<= 2 pages in English) showing your procedure and results (i.e., testing accuracy). The use of randome forest is very simple. If train and test are dataframes with features x and class label y, you use

myrf<-randomForest(as.matrix(train$x), train$y, as.matrix(test$x), test$y)
More information can be found its manual (also availablem from CRAN).
Last modified: Fri Feb 13 06:17:14 CST 2004