From d345177ce98071a6c5aa93fec126b2b987ab4ebe Mon Sep 17 00:00:00 2001
From: Aaron Keesing <agkphysics@gmail.com>
Date: Mon, 26 Aug 2024 00:15:13 +1200
Subject: [PATCH] Change default parameters. Check parameter values during
 training.

---
 .../classifiers/multilabel/meta/MLRF.java     | 39 ++++++++++++-------
 1 file changed, 25 insertions(+), 14 deletions(-)

diff --git a/src/main/java/meka/classifiers/multilabel/meta/MLRF.java b/src/main/java/meka/classifiers/multilabel/meta/MLRF.java
index 9601401a..1a8571f2 100644
--- a/src/main/java/meka/classifiers/multilabel/meta/MLRF.java
+++ b/src/main/java/meka/classifiers/multilabel/meta/MLRF.java
@@ -32,6 +32,7 @@
 import no.uib.cipr.matrix.SVD;
 import no.uib.cipr.matrix.sparse.CompRowMatrix;
 import weka.core.Attribute;
+import weka.core.Capabilities;
 import weka.core.Instance;
 import weka.core.Instances;
 import weka.core.Option;
@@ -40,9 +41,9 @@
 import weka.core.TechnicalInformation.Field;
 import weka.core.TechnicalInformation.Type;
 import weka.core.TechnicalInformationHandler;
+import weka.core.WekaException;
 
 import java.util.ArrayList;
-import java.util.Collections;
 import java.util.Enumeration;
 import java.util.List;
 import java.util.Random;
@@ -60,12 +61,12 @@ public class MLRF extends MetaProblemTransformationMethod implements Randomizabl
 	/**
 	 * The number of subsets to generate.
 	 */
-	protected int K = 10;
+	protected int K = 2;
 
 	/**
 	 * The dimensionality reduction parameter.
 	 */
-	protected int numFeatures = 10;
+	protected int numFeatures = 5;
 
 	private DenseMatrix[] R;
 
@@ -121,14 +122,12 @@ private List<List<Integer>> generateFeatureSubsets(int d) {
 		A.shuffle(indices, r);
 		for (int i = 0; i < K; i++) {
 			List<Integer> subset = new ArrayList<>();
-			for (int j = 0; j < d / K; j++) {
-				if (i * (d / K) + j >= d)
-					break;
+			for (int j = 0; j < d / K; j++)
 				subset.add(indices[i * (d / K) + j]);
-			}
-			Collections.sort(subset);
 			result.add(subset);
 		}
+		for (int i = K * (d / K); i < d; i++)
+			result.get(i - K * (d / K)).add(indices[i]);
 		return result;
 	}
 
@@ -139,6 +138,11 @@ public void buildClassifier(Instances D) throws Exception {
 		int L = D.classIndex();
 		int d = D.numAttributes() - L;
 
+		if (K > d)
+			throw new WekaException("Number of subsets is larger than the number of attributes.");
+		if (numFeatures > (d / K))
+			throw new WekaException("Dimensionality reduction parameter k is larger than the subset size.");
+
 		m_Classifiers = ProblemTransformationMethod.makeCopies((MultiLabelClassifier) m_Classifier, m_NumIterations);
 		R = new DenseMatrix[m_NumIterations];
 		Instances transformedInstances = F.remove(D, A.make_sequence(L), true);
@@ -195,8 +199,7 @@ public void buildClassifier(Instances D) throws Exception {
 				}
 
 				// Latent semantic indexing
-				SVD svd = new SVD(nNew, subSize);
-				svd = svd.factor(mat);
+				SVD svd = SVD.factorize(mat);
 				DenseMatrix transformationMatrix = getTransformationMatrix(svd, subSize);
 
 				// Insert each subset transformation matrix into full rotation matrix R[i]
@@ -279,8 +282,8 @@ public String[] getOptions() {
 
 	@Override
 	public void setOptions(String[] options) throws Exception {
-		K = OptionUtils.parse(options, "K", 10);
-		numFeatures = OptionUtils.parse(options, "k", 10);
+		K = OptionUtils.parse(options, "K", 2);
+		numFeatures = OptionUtils.parse(options, "k", 5);
 		super.setOptions(options);
 	}
 
@@ -307,7 +310,15 @@ public TechnicalInformation getTechnicalInformation() {
 		return info;
 	}
 
-    public static void main(String[] args) {
-        ProblemTransformationMethod.runClassifier(new MLRF(), args);
+	@Override
+	public Capabilities getCapabilities() {
+		Capabilities result = super.getCapabilities();
+		result.disable(Capabilities.Capability.NOMINAL_ATTRIBUTES);
+		result.disable(Capabilities.Capability.MISSING_VALUES);
+		return result;
+	}
+
+	public static void main(String[] args) {
+		ProblemTransformationMethod.runClassifier(new MLRF(), args);
     }
 }