Fixed evaluate method for binary classification

Author: xEcho1337

src/main/java/net/echo/brain4j/model/impl/Sequential.java

@@ -103,7 +103,12 @@ public Sequential compile(WeightInit weightInit, LossFunctions function, Optimiz
 
     @Override
     public EvaluationResult evaluate(DataSet<DataRow> dataSet) {
-        int classes = layers.getLast().getNeurons().size();
+        int classes = dataSet.getData().getFirst().outputs().size();
+
+        // Binary classification
+        if (classes == 1) {
+            classes = 2;
+        }
 
         Map<Integer, Vector> classifications = new ConcurrentHashMap<>();
 
@@ -133,6 +138,11 @@ private Thread makeEvaluation(List<DataRow> partition, Map<Integer, Vector> clas
                 int predIndex = MLUtils.indexOfMaxValue(prediction);
                 int targetIndex = MLUtils.indexOfMaxValue(row.outputs());
 
+                if (row.outputs().size() == 1) {
+                    predIndex = prediction.get(0) > 0.5 ? 1 : 0;
+                    targetIndex = (int) row.outputs().get(0);
+                }
+
                 Vector predictions = classifications.get(targetIndex);
                 predictions.set(predIndex, predictions.get(predIndex) + 1);
             }

src/main/java/net/echo/brain4j/training/evaluation/EvaluationResult.java

@@ -74,7 +74,7 @@ public String confusionMatrix() {
         }
 
         matrix.append("\n  ");
-        matrix.append("-".repeat(classes * 6)).append("\n");
+        matrix.append("-".repeat(5 + classes * 5)).append("\n");
 
         for (int i = 0; i < classes; i++) {
             StringBuilder text = new StringBuilder();

src/test/java/XorExample.java

@@ -23,22 +23,15 @@ private void start() {
         System.out.println(model.getStats());
 
         long start = System.nanoTime();
-
-        // Fit the model for 1000 epoches
-        for (int i = 0; i < 100; i++) {
-            model.fit(dataSet);
-        }
-
+        model.fit(dataSet, 1000);
         double took = (System.nanoTime() - start) / 1e6;
 
         System.out.println("Loss: " + model.loss(dataSet));
         System.out.println("Took: " + took + " ms");
 
-        for (DataRow row : dataSet) {
-            Vector prediction = model.predict(row.inputs());
+        var result = model.evaluate(dataSet);
 
-            System.out.println("Expected: " + row.outputs() + " -> Predicted: " + prediction.toString("%.3f"));
-        }
+        System.out.println(result.confusionMatrix());
     }
 
     private Sequential getModel() {