Java: Data Science Made Easy

JSAT

JSAT is also a general purpose library and contains a lot of implementations for solving regression problems.

As with classification, it needs a wrapper class for data and a special wrapper for regression:

double[][] X = ... // double[] y = ... // List<DataPointPair<Double>> data = new ArrayList<>(X.length); for (int i = 0; i < X.length; i++) {     DataPoint row = new DataPoint(new DenseVector(X[i]));     data.add(new DataPointPair<Double>(row, y[i])); } RegressionDataSet dataset = new RegressionDataSet(data);

Once the dataset is wrapped in the right class, we can train models like this:

MultipleLinearRegression linreg = new MultipleLinearRegression(); linreg.train(dataset);;

The preceding code trains the usual OLS linear regression.