This library is made to simplify the use of regression on machine learning
- jblas-1.2.4
- java 8
At the moment, there is only one way available to use the API.
You can build the project from the source in this repository, export as a JAR file and Add to the Build Path of your project.
// Here, we load the data necessary to train the model.
DoubleMatrix dataset = LoadData.load("train_data/data_1.txt", ",");
// split the data in two: features and expected answers for each set of features(row)
DoubleMatrix X = dataset.getColumn(0);
DoubleMatrix Y = dataset.getColumn(1);
/*define the model to use. Obs: first parameter is the learning coeficient,
*second parameter is the number of iterations to
*train the model e finally the last parameter is a question "Do you want to normalize the data?".
*/
Regression model = new LinearRegression(0.01, 2000, false);
model.train(X, Y);
// After training the model we predict some values, just to test the model
DoubleMatrix predict1 = model.predict(new DoubleMatrix(new double[] {3.5}));
DoubleMatrix predict2 = model.predict(new DoubleMatrix(new double[] {7}));
System.out.println("For population = 35,000, we predict a profit of " + (predict1.get(0) * 10000));
System.out.println("For population = 70,000, we predict a profit of " + (predict2.get(0) * 10000));
```