Changeset 9 for trunk/EventBenchCore/src/de/ugoe/cs/eventbench/ppm/Trie.java

Timestamp:

04/13/11 14:13:45 (13 years ago)

Author:

sherbold

Message:

• moved training logic for single sequences from PPM to Trie

File:

• trunk/EventBenchCore/src/de/ugoe/cs/eventbench/ppm/Trie.java (modified) (2 diffs)

trunk/EventBenchCore/src/de/ugoe/cs/eventbench/ppm/Trie.java

@@ -4 +4 @@
 import java.awt.Rectangle;
 import java.awt.Shape;
+import java.util.LinkedHashSet;
 import java.util.LinkedList;
 import java.util.List;
+import java.util.Set;
 
 import javax.swing.JFrame;
 
 import org.apache.commons.collections15.Transformer;
+
+import de.ugoe.cs.eventbench.markov.IncompleteMemory;
 
 import edu.uci.ics.jung.algorithms.layout.Layout;
@@ -21 +25 @@
 public class Trie<T> {
         
+        private Set<T> knownSymbols;
+        
         private final TrieNode<T> rootNode;
         
         public Trie() {
                 rootNode = new TrieNode<T>();
+                knownSymbols = new LinkedHashSet<T>();
+        }
+        
+        public Set<T> getKnownSymbols() {
+                return new LinkedHashSet<T>(knownSymbols);
+        }
+        
+        // trains the current Trie using the given sequence and adds all subsequence of length maxOrder
+        public void train(List<T> sequence, int maxOrder) {
+                IncompleteMemory<T> latestActions = new IncompleteMemory<T>(maxOrder);
+                int i=0;
+                for(T currentEvent : sequence) {
+                        latestActions.add(currentEvent);
+                        knownSymbols.add(currentEvent);
+                        i++;
+                        if( i>=maxOrder ) {
+                                add(latestActions.getLast(maxOrder));
+                        }
+                }
+                int sequenceLength = sequence.size();
+                for( int j=maxOrder-1 ; j>0 ; j-- ) {
+                        add(sequence.subList(sequenceLength-j, sequenceLength));
+                }
         }
         
 
+        // increases the counters for each symbol in the subsequence
         public void add(List<T> subsequence) {
-                if( !subsequence.isEmpty() ) {
+                if( subsequence!=null && !subsequence.isEmpty() ) {
+                        knownSymbols.addAll(subsequence);
                         subsequence = new LinkedList<T>(subsequence);  // defensive copy!
                         T firstSymbol = subsequence.get(0);