June 2017
Beginner to intermediate
576 pages
15h 22m
English
Aside from the prediction accuracy, another popular use for random forests is variable selection, using the varImp() function. A variable importance plot can be useful in situations in which there are many input variables, but I have found it to be of limited value for a manageable number of variables.
Just to illustrate on our example data, here is the varImp() function showing Fare, Age, and Pclass in the order of importance. There are a couple of ways it can do this, but I will be showing it as determined via statistics referred to as MeanDecreaseGini. It is not critical to understand how this statistic is computed, but it is sufficient to say at this point that the importance is related to how many different ...