Add LearningDataSet class

src/main/java/net/zomis/machlearn/common/ClassifierFunction.java

@@ -0,0 +1,7 @@
+package net.zomis.machlearn.common;
+
+public interface ClassifierFunction {
+
+    boolean classify(double[] theta, double[] x);
+
+}

src/main/java/net/zomis/machlearn/common/LearningDataSet.java

@@ -0,0 +1,51 @@
+package net.zomis.machlearn.common;
+
+import net.zomis.machlearn.neural.LearningData;
+import net.zomis.machlearn.text.duga.PrecisionRecallF1;
+
+import java.util.ArrayList;
+import java.util.List;
+import java.util.stream.Collectors;
+
+public class LearningDataSet {
+
+    private final List<LearningData> data = new ArrayList<>();
+
+    public void add(Object representation, double[] x, double y) {
+        this.add(representation, x, new double[]{y});
+    }
+
+    public void add(Object representation, double[] x, double[] y) {
+        this.data.add(new LearningData(representation, x, y));
+    }
+
+    public double[][] getXs() {
+        return data.stream()
+            .map(LearningData::getInputs)
+            .collect(Collectors.toList()).toArray(new double[data.size()][]);
+    }
+
+    public double[] getY() {
+        return data.stream()
+            .map(LearningData::getOutputs)
+            .mapToDouble(d -> d[0]).toArray();
+    }
+
+    public int numFeaturesWithZero() {
+        return data.get(0).getInputs().length + 1;
+    }
+
+    public PrecisionRecallF1 precisionRecallF1(double[] theta, ClassifierFunction hypothesis) {
+        PrecisionRecallF1 score = new PrecisionRecallF1();
+        for (LearningData ld : data) {
+            boolean prediction = hypothesis.classify(theta, ld.getInputs());
+            boolean actual = ld.getOutputs()[0] >= 0.5;
+            score.add(actual, prediction);
+        }
+        return score;
+    }
+
+    public List<LearningData> getData() {
+        return data;
+    }
+}