i really like your code, works nice, readable, does exactly, what i was looking for, etc, but:

if i just try(naively):

go get "github.com/fxsjy/RF.go"

it's pulling in the whole 47mb mnist database, among other goodies there too ...

I'm following the iris example and built a similar forest for malware detection. My inputs is a vector of ints and the target a string.

However when I use forest.Predicate(inputs) I don't get any output (empty string).. Is it something intended?

Here is a snippet of my code :

        // split train and test data
	Logger.Debug("splitting data in training and testing")
	train_input := [][]interface{}{}
	test_input := [][]interface{}{}

	train_target := []string{}
	test_target := []string{}

	for i := 0; i < len(inputs); i++ {
		input := inputs[i]
		target := targets[i]

		test := rand.Float64()
		if test > TRAIN_TEST_PERCENT {
			test_input = append(test_input, input)
			test_target = append(test_target, target)
		} else {
			train_input = append(train_input, input)
			train_target = append(train_target, target)
		}
	}
	Logger.Debugf("Got %d training and %d testing samples over %d samples", len(train_input), len(test_input), len(inputs))

	// create classifier
	Logger.Debug("training the forest")
	forest := RF.DefaultForest(train_input, train_target, 100) //100 trees
	RF.DumpForest(forest, filepath.Join("data", "model.ml"))

	// get stats for the classifier
	Logger.Debug("gathering stats")
	err_count := 0.0
	for i := 0; i < len(test_input); i++ {
		output := forest.Predicate(test_input[i])
		expect := test_target[i]
		Logger.Tracef("Out: %s | Expected: %s", output, expect)
		if output != expect {
			err_count += 1
		}
	}
	Logger.Infof("success rate: %f", 1.0-err_count/float64(len(test_input)))

Hi, we want to use your awesome library in production, but we need to know exactly what license it is under. There is no LICENSE file or license header in source files. Could you please clarify it? Thanks!