diff --git a/src/main/java/be/cylab/java/wowa/training/Example.java b/src/main/java/be/cylab/java/wowa/training/Example.java
index f601d3e466e3f1d382a687fd36bb18f0cb289a41..d1055c0ba01da1379640af927d2298bbb5edc1e3 100644
--- a/src/main/java/be/cylab/java/wowa/training/Example.java
+++ b/src/main/java/be/cylab/java/wowa/training/Example.java
@@ -1,6 +1,8 @@
package be.cylab.java.wowa.training;
+import java.util.HashMap;
+import java.util.Map;
import java.util.logging.Level;
import java.util.logging.Logger;
import java.util.logging.SimpleFormatter;
@@ -72,15 +74,15 @@ public final class Example {
+ (end_time - start_time) / 1000 + " seconds");
System.out.println("Run Cross validation");
- //HashMap<AbstractSolution, Double> solutions = trainer.runKFold(
- // data_file,
- // expected_file,
- // 10,
- // 10);
+ HashMap<AbstractSolution, Double> solutions = trainer.runKFold(
+ data_file,
+ expected_file,
+ 10,
+ 10);
- //for (Map.Entry val : solutions.entrySet()) {
- // System.out.println(val);
- //}
+ for (Map.Entry val : solutions.entrySet()) {
+ System.out.println(val);
+ }
}
diff --git a/src/main/java/be/cylab/java/wowa/training/Trainer.java b/src/main/java/be/cylab/java/wowa/training/Trainer.java
index 01ff230160e5c5146e702b8f4c128f05b6b50000..d53397826145c2fe7d4691070644f00d80dbe77c 100644
--- a/src/main/java/be/cylab/java/wowa/training/Trainer.java
+++ b/src/main/java/be/cylab/java/wowa/training/Trainer.java
@@ -125,6 +125,12 @@ public class Trainer {
testing.getData(),
testing.getExpected());
+ Utils.computeStatisticalInformation(
+ testing.getData(),
+ testing.getExpected(),
+ 100,
+ "Fold_" + (i + 1) + ".txt",
+ sol);
map.put(sol, score);
}
return map;
diff --git a/src/main/java/be/cylab/java/wowa/training/Utils.java b/src/main/java/be/cylab/java/wowa/training/Utils.java
index 362b56ec110d0910d96e3554caea837ae2975ee3..2ebce94788805155e82acbffc2e8bb00286c1295 100644
--- a/src/main/java/be/cylab/java/wowa/training/Utils.java
+++ b/src/main/java/be/cylab/java/wowa/training/Utils.java
@@ -2,9 +2,10 @@ package be.cylab.java.wowa.training;
import com.owlike.genson.GenericType;
import com.owlike.genson.Genson;
+import info.debatty.java.aggregation.WOWA;
import org.apache.commons.lang3.ArrayUtils;
-import java.io.IOException;
+import java.io.*;
import java.nio.charset.StandardCharsets;
import java.nio.file.Files;
import java.nio.file.Paths;
@@ -242,4 +243,175 @@ final class Utils {
}
}
+ static void computeStatisticalInformation(
+ final List<List<Double>> data,
+ final List<Double> expected,
+ final int nbr_values,
+ final String filename,
+ final AbstractSolution solution) {
+ int wowa_true_positive_counter = 0;
+ int wowa_true_negative_counter = 0;
+ int average_true_positive_counter = 0;
+ int average_true_negative_counter = 0;
+
+ int wowa_false_positive_counter = 0;
+ int wowa_false_negative_counter = 0;
+ int average_false_positive_counter = 0;
+ int average_false_negative_counter = 0;
+
+ double iteration = 1 / (double) nbr_values;
+
+
+ for (double trigger_value = 0.00;
+ trigger_value < 1; trigger_value += iteration) {
+ for (int j = 0; j < data.size(); j++) {
+ WOWA wowa = new WOWA(Utils.convertListDoubleToArrayDouble(
+ solution.weights_w),
+ Utils.convertListDoubleToArrayDouble(
+ solution.weights_p));
+ double wowa_result = wowa.aggregate(
+ Utils.convertListDoubleToArrayDouble(data.get(j)));
+ double average = Utils.sumListElements(data.get(j))
+ / data.get(j).size();
+ if (expected.get(j) == 1 && average > trigger_value && wowa_result > trigger_value) {
+ average_true_positive_counter++;
+ wowa_true_positive_counter++;
+ } else if (expected.get(j) == 1 && average < trigger_value && wowa_result > trigger_value) {
+ wowa_true_positive_counter++;
+ average_false_negative_counter++;
+ } else if (expected.get(j) == 1 && average > trigger_value && wowa_result < trigger_value) {
+ average_true_positive_counter++;
+ wowa_false_negative_counter++;
+ } else if (expected.get(j) == 1 && average < trigger_value && wowa_result < trigger_value) {
+ average_false_negative_counter++;
+ wowa_false_negative_counter++;
+ } else if (expected.get(j) == 0 && average < trigger_value && wowa_result < trigger_value) {
+ average_true_negative_counter++;
+ wowa_true_negative_counter++;
+ } else if (expected.get(j) == 0 && average < trigger_value && wowa_result > trigger_value) {
+ average_true_negative_counter++;
+ wowa_false_positive_counter++;
+ } else if (expected.get(j) == 0 && average > trigger_value && wowa_result < trigger_value) {
+ wowa_true_negative_counter++;
+ average_false_positive_counter++;
+ } else if (expected.get(j) == 0 && average > trigger_value && wowa_result > trigger_value) {
+ average_false_positive_counter++;
+ wowa_false_positive_counter++;
+ }
+ }
+
+ //Positive conditions.
+ int wowa_positive_condition = wowa_true_positive_counter + wowa_false_negative_counter;
+ int average_positive_condition = average_true_positive_counter + average_false_negative_counter;
+
+ //Negative conditions.
+ int wowa_negative_condition = wowa_true_negative_counter + wowa_false_positive_counter;
+ int average_negative_condition = average_true_negative_counter + average_false_positive_counter;
+
+ //True positive rate.
+ double wowa_true_positive_rate = wowa_true_positive_counter / wowa_positive_condition;
+ double average_true_positive_rate = average_true_positive_counter / average_positive_condition;
+
+ //True negative rate.
+ double wowa_true_negative_rate = wowa_true_negative_counter / wowa_negative_condition;
+ double average_true_negative_rate = average_true_negative_counter / average_negative_condition;
+
+ //False negative rate.
+ double wowa_false_negative_rate = 1 - wowa_true_positive_rate;
+ double average_false_negative_rate = 1 - average_true_positive_rate;
+
+ //Flase positive rate.
+ double wowa_false_positive_rate = 1 - wowa_true_negative_rate;
+ double average_false_positive_rate = 1 - average_true_negative_rate;
+
+ //Precision (or positive prediction value).
+ double wowa_precision = wowa_true_positive_counter / (double) (wowa_true_positive_counter + wowa_false_positive_counter);
+ double average_precision = average_true_positive_counter / (double) (average_true_positive_counter + average_false_positive_counter);
+
+ //Negative prediction value.
+ double wowa_negative_prediction_value = wowa_true_negative_counter / (double) (wowa_true_negative_counter + wowa_false_negative_counter);
+ double average_negative_prediction_value = average_true_negative_counter / (double) (average_true_negative_counter + average_false_negative_counter);
+
+ //False discovery rate.
+ double wowa_false_discovery_rate = 1 - wowa_precision;
+ double average_false_discovery_rate = 1 - average_precision;
+
+ //False omission rate.
+ double wowa_false_omission_rate = 1 - wowa_negative_prediction_value;
+ double average_false_omission_rate = 1 - average_negative_prediction_value;
+
+ //Accuracy.
+ double wowa_accuracy = (wowa_true_positive_counter + wowa_true_negative_counter) / expected.size();
+ double average_accuracy = (average_true_positive_counter + average_true_negative_counter) / expected.size();
+
+ //F1 score.
+ double wowa_f_1 = 2 * (wowa_precision * wowa_true_positive_rate) / (wowa_precision + wowa_true_positive_rate);
+ double average_f_1 = 2 * (average_precision * average_true_positive_rate) / (average_precision + average_true_positive_rate);
+
+ try (FileWriter fw = new FileWriter(filename, true);
+ BufferedWriter bw = new BufferedWriter(fw);
+ PrintWriter writer = new PrintWriter(bw)) {
+
+ writer.println("___________________________________________");
+ writer.println("|TRIGGER VALUE :" + trigger_value);
+ writer.println("|_________________________________________|");
+ writer.println("WOWA");
+
+ writer.println("Wowa true positive : " + wowa_true_positive_counter);
+ writer.println("Wowa true negative : " + wowa_true_negative_counter);
+ writer.println("Wowa false positive : " + wowa_false_positive_counter);
+ writer.println("Wowa false negative : " + wowa_false_negative_counter);
+
+ writer.println("Wowa true positive rate : " + wowa_true_positive_rate);
+ writer.println("Wowa true negative rate : " + wowa_true_negative_rate);
+ writer.println("Wowa false positive rate : " + wowa_false_positive_rate);
+ writer.println("Wowa false negative rate : " + wowa_false_negative_rate);
+
+ writer.println("Wowa precision : " + wowa_precision);
+ writer.println("Wowa negative predictive value : " + wowa_negative_prediction_value);
+ writer.println("Wowa false discovery rate : " + wowa_false_discovery_rate);
+ writer.println("Wowa false omission rate : " + wowa_false_omission_rate);
+ writer.println("Wowa accuracy : " + wowa_accuracy);
+ writer.println("Wowa F1 score : " + wowa_f_1);
+
+ writer.println("AVERAGE\n");
+ writer.println("Average true positive : " + average_true_positive_counter);
+ writer.println("Average true negative : " + average_true_negative_counter);
+ writer.println("Average false positive : " + average_false_positive_counter);
+ writer.println("Average false negative : " + average_false_negative_counter);
+
+ writer.println("Average true positive rate : " + average_true_positive_rate);
+ writer.println("Average true negative rate : " + average_true_negative_rate);
+ writer.println("Average false positive rate : " + average_false_positive_rate);
+ writer.println("Average false negative rate : " + average_false_negative_rate);
+
+ writer.println("Average precision : " + average_precision);
+ writer.println("Average negative predictive value : " + average_negative_prediction_value);
+ writer.println("Average false discovery rate : " + average_false_discovery_rate);
+ writer.println("Average false omission rate : " + average_false_omission_rate);
+ writer.println("Average accuracy : " + average_accuracy);
+ writer.println("Average F1 score : " + average_f_1);
+
+ wowa_true_positive_counter = 0;
+ wowa_true_negative_counter = 0;
+ average_true_positive_counter = 0;
+ average_true_negative_counter = 0;
+
+ wowa_false_positive_counter = 0;
+ wowa_false_negative_counter = 0;
+ average_false_positive_counter = 0;
+ average_false_negative_counter = 0;
+
+ } catch (FileNotFoundException e) {
+ e.printStackTrace();
+ } catch (UnsupportedEncodingException e) {
+ e.printStackTrace();
+ } catch (IOException e) {
+ e.printStackTrace();
+ }
+
+ }
+
+ }
+
}