From 368af15fd8368cf6694200f1715a0468083740d0 Mon Sep 17 00:00:00 2001
From: joethei <j.theiner@live.de>
Date: Thu, 25 Apr 2019 13:02:59 +0200
Subject: [PATCH] =?UTF-8?q?Cluster=20fast=20fertig,=20ein=20bischen=20aufr?=
 =?UTF-8?q?=C3=A4umen=20noch=20n=C3=B6tig?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

#SPM-13: add work development 2h 20m Cluster fertig
---
 .../de/hsel/spm/baudas/analysis/Cluster.java  | 88 +++++++++++++++++++
 .../hsel/spm/baudas/analysis/ClusterTest.java | 29 ++++++
 2 files changed, 117 insertions(+)
 create mode 100644 src/main/java/de/hsel/spm/baudas/analysis/Cluster.java
 create mode 100644 src/test/java/de/hsel/spm/baudas/analysis/ClusterTest.java

diff --git a/src/main/java/de/hsel/spm/baudas/analysis/Cluster.java b/src/main/java/de/hsel/spm/baudas/analysis/Cluster.java
new file mode 100644
index 0000000..d340747
--- /dev/null
+++ b/src/main/java/de/hsel/spm/baudas/analysis/Cluster.java
@@ -0,0 +1,88 @@
+package de.hsel.spm.baudas.analysis;
+
+import weka.clusterers.SimpleKMeans;
+import weka.core.Instances;
+import weka.filters.Filter;
+import weka.filters.unsupervised.attribute.Remove;
+
+import java.io.File;
+import java.util.HashMap;
+import java.util.Map;
+
+/**
+ * Cluster Analysis
+ *
+ * @author Johannes Theiner
+ * @version 0.1
+ */
+public class Cluster implements Analysis<Map<Integer, Map<String, String>>> {
+
+    private Instances instances;
+    private Map<Integer, Map<String, String>> result;
+
+    public Cluster(File file) {
+        instances = load(file);
+    }
+
+    @Override
+    public Map<Integer, Map<String, String>> getResult() {
+        if(result == null) {
+            result = new HashMap<>();
+            //TODO: anpassen wenn #SPM-17 gemerged ist.
+            int[] keepIndexes = new int[]{0, 1, 3, 5, 6, 7};
+            Remove remove = new Remove();
+
+            try {
+                remove.setAttributeIndicesArray(keepIndexes);
+                remove.setInvertSelection(true);
+                remove.setInputFormat(instances);
+                instances = Filter.useFilter(instances, remove);
+            } catch (Exception e) {
+                e.printStackTrace();
+            }
+
+            SimpleKMeans fullMeans = new SimpleKMeans();
+            Instances fullCentroids = null;
+            try {
+                fullMeans.setNumClusters(1);
+                fullMeans.setPreserveInstancesOrder(true);
+                fullMeans.buildClusterer(instances);
+                fullCentroids = fullMeans.getClusterCentroids();
+
+            } catch (Exception ex) {
+                ex.printStackTrace();
+            }
+            assert fullCentroids != null;
+
+            SimpleKMeans kMeans = new SimpleKMeans();
+            try {
+                kMeans.setNumClusters(5);
+                kMeans.setPreserveInstancesOrder(true);
+                kMeans.buildClusterer(instances);
+
+                int count = 0;
+                count = putIntoMap(fullCentroids, count);
+
+                Instances centroids = kMeans.getClusterCentroids();
+                putIntoMap(centroids, count);
+
+            } catch (Exception ex) {
+                ex.printStackTrace();
+            }
+        }
+
+        return result;
+    }
+
+    private int putIntoMap(Instances centroids, int count) {
+        for (int i = 0; i < centroids.numInstances(); i++) {
+            Map<String, String> map = new HashMap<>();
+            for (int j = 0; j < centroids.numAttributes(); j++) {
+                map.put(centroids.attribute(j).name(), centroids.instance(i).stringValue(j));
+            }
+            result.put(count, map);
+            count++;
+        }
+        return count;
+    }
+}
\ No newline at end of file
diff --git a/src/test/java/de/hsel/spm/baudas/analysis/ClusterTest.java b/src/test/java/de/hsel/spm/baudas/analysis/ClusterTest.java
new file mode 100644
index 0000000..2711fa6
--- /dev/null
+++ b/src/test/java/de/hsel/spm/baudas/analysis/ClusterTest.java
@@ -0,0 +1,29 @@
+package de.hsel.spm.baudas.analysis;
+
+import org.junit.jupiter.api.Test;
+
+import java.io.File;
+import java.net.URL;
+import java.time.Duration;
+import java.util.Map;
+import java.util.concurrent.atomic.AtomicReference;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertTimeout;
+
+public class ClusterTest {
+
+    @Test
+    public void test100() {
+        URL url = getClass().getClassLoader().getResource("kd100.csv");
+        assert url != null;
+
+        Cluster cluster = new Cluster(new File(url.getFile()));
+
+        AtomicReference<Map<Integer, Map<String, String>>> results = new AtomicReference<>();
+
+        assertTimeout(Duration.ofMillis(2000), () -> results.set(cluster.getResult()));
+
+        assertEquals("m", results.get().get(0).get("Geschlecht"));
+    }
+}
\ No newline at end of file