Machine Learning With Go

Decision tree example

We will again utilize the iris dataset for this example. You have already learned how to handle this dataset in github.com/sjwhitworth/golearn, and we can follow a similar pattern again. We will again use cross-validation. However, this time we will fit a decision tree model:

// Read in the iris data set into golearn "instances". irisData, err := base.ParseCSVToInstances("iris.csv", true) if err != nil { log.Fatal(err) } // This is to seed the random processes involved in building the // decision tree. rand.Seed(44111342) // We will use the ID3 algorithm to build our decision tree. Also, we // will start with a parameter of 0.6 that controls the train-prune split. tree := trees.NewID3DecisionTree(0.6) // Use cross-fold ...