source: trunk/EventBenchCore/src/de/ugoe/cs/eventbench/models/TrieBasedModel.java @ 100

package de.ugoe.cs.eventbench.models;

import java.security.InvalidParameterException;
import java.util.ArrayList;
import java.util.LinkedHashSet;
import java.util.LinkedList;
import java.util.List;
import java.util.Random;
import java.util.Set;

import de.ugoe.cs.eventbench.data.Event;
import de.ugoe.cs.eventbench.models.Trie.Edge;
import de.ugoe.cs.eventbench.models.Trie.TrieVertex;
import edu.uci.ics.jung.graph.Tree;

/**
 * <p>
 * Implements a skeleton for stochastic processes that can calculate
 * probabilities based on a trie. The skeleton provides all functionalities of
 * {@link IStochasticProcess} except
 * {@link IStochasticProcess#getProbability(List, Event)}.
 * </p>
 * 
 * @author Steffen Herbold
 * @version 1.0
 */
public abstract class TrieBasedModel implements IStochasticProcess {

        /**
         * <p>
         * Id for object serialization.
         * </p>
         */
        private static final long serialVersionUID = 1L;

        /**
         * <p>
         * The order of the trie, i.e., the maximum length of subsequences stored in
         * the trie.
         * </p>
         */
        protected int trieOrder;

        /**
         * <p>
         * Trie on which the probability calculations are based.
         * </p>
         */
        protected Trie<Event<?>> trie;

        /**
         * <p>
         * Random number generator used by probabilistic sequence generation
         * methods.
         * </p>
         */
        protected final Random r;

        /**
         * <p>
         * Constructor. Creates a new TrieBasedModel that can be used for stochastic
         * processes with a Markov order less than or equal to {@code markovOrder}.
         * </p>
         * 
         * @param markovOrder
         *            Markov order of the model
         * @param r
         *            random number generator used by probabilistic methods of the
         *            class
         */
        public TrieBasedModel(int markovOrder, Random r) {
                super();
                this.trieOrder = markovOrder + 1;
                this.r = r;
        }

        /**
         * <p>
         * Trains the model by generating a trie from which probabilities are
         * calculated.
         * </p>
         * 
         * @param sequences
         *            training data
         */
        public void train(List<List<Event<?>>> sequences) {
                trie = new Trie<Event<?>>();

                for (List<Event<?>> sequence : sequences) {
                        List<Event<?>> currentSequence = new LinkedList<Event<?>>(sequence); // defensive
                                                                                                                                                                        // copy
                        currentSequence.add(0, Event.STARTEVENT);
                        currentSequence.add(Event.ENDEVENT);

                        trie.train(currentSequence, trieOrder);
                }
        }

        /*
         * (non-Javadoc)
         * 
         * @see de.ugoe.cs.eventbench.models.IStochasticProcess#randomSequence()
         */
        @Override
        public List<? extends Event<?>> randomSequence() {
                List<Event<?>> sequence = new LinkedList<Event<?>>();

                IncompleteMemory<Event<?>> context = new IncompleteMemory<Event<?>>(
                                trieOrder - 1);
                context.add(Event.STARTEVENT);

                Event<?> currentState = Event.STARTEVENT;

                boolean endFound = false;

                while (!endFound) {
                        double randVal = r.nextDouble();
                        double probSum = 0.0;
                        List<Event<?>> currentContext = context.getLast(trieOrder);
                        for (Event<?> symbol : trie.getKnownSymbols()) {
                                probSum += getProbability(currentContext, symbol);
                                if (probSum >= randVal) {
                                        endFound = (symbol == Event.ENDEVENT);
                                        if (!(symbol == Event.STARTEVENT || symbol == Event.ENDEVENT)) {
                                                // only add the symbol the sequence if it is not START
                                                // or END
                                                context.add(symbol);
                                                currentState = symbol;
                                                sequence.add(currentState);
                                        }
                                        break;
                                }
                        }
                }
                return sequence;
        }

        /**
         * <p>
         * Returns a Dot representation of the internal trie.
         * </p>
         * 
         * @return dot representation of the internal trie
         */
        public String getTrieDotRepresentation() {
                return trie.getDotRepresentation();
        }

        /**
         * <p>
         * Returns a {@link Tree} of the internal trie that can be used for
         * visualization.
         * </p>
         * 
         * @return {@link Tree} depicting the internal trie
         */
        public Tree<TrieVertex, Edge> getTrieGraph() {
                return trie.getGraph();
        }

        /**
         * <p>
         * The string representation of the model is {@link Trie#toString()} of
         * {@link #trie}.
         * </p>
         * 
         * @see java.lang.Object#toString()
         */
        @Override
        public String toString() {
                return trie.toString();
        }

        /*
         * (non-Javadoc)
         * 
         * @see de.ugoe.cs.eventbench.models.IStochasticProcess#getNumStates()
         */
        @Override
        public int getNumStates() {
                return trie.getNumSymbols();
        }

        /*
         * (non-Javadoc)
         * 
         * @see de.ugoe.cs.eventbench.models.IStochasticProcess#getStateStrings()
         */
        @Override
        public String[] getStateStrings() {
                String[] stateStrings = new String[getNumStates()];
                int i = 0;
                for (Event<?> symbol : trie.getKnownSymbols()) {
                        stateStrings[i] = symbol.toString();
                        i++;
                }
                return stateStrings;
        }

        /*
         * (non-Javadoc)
         * 
         * @see de.ugoe.cs.eventbench.models.IStochasticProcess#getEvents()
         */
        @Override
        public Set<? extends Event<?>> getEvents() {
                return trie.getKnownSymbols();
        }

        /*
         * (non-Javadoc)
         * 
         * @see
         * de.ugoe.cs.eventbench.models.IStochasticProcess#generateSequences(int)
         */
        @Override
        public Set<List<? extends Event<?>>> generateSequences(int length) {
                return generateSequences(length, false);
        }

        /*
         * (non-Javadoc)
         * 
         * @see
         * de.ugoe.cs.eventbench.models.IStochasticProcess#generateSequences(int,
         * boolean)
         */
        @Override
        public Set<List<? extends Event<?>>> generateSequences(int length,
                        boolean fromStart) {
                Set<List<? extends Event<?>>> sequenceSet = new LinkedHashSet<List<? extends Event<?>>>();
                ;
                if (length < 1) {
                        throw new InvalidParameterException(
                                        "Length of generated subsequences must be at least 1.");
                }
                if (length == 1) {
                        if (fromStart) {
                                List<Event<?>> subSeq = new LinkedList<Event<?>>();
                                subSeq.add(Event.STARTEVENT);
                                sequenceSet.add(subSeq);
                        } else {
                                for (Event<?> event : getEvents()) {
                                        List<Event<?>> subSeq = new LinkedList<Event<?>>();
                                        subSeq.add(event);
                                        sequenceSet.add(subSeq);
                                }
                        }
                        return sequenceSet;
                }
                Set<? extends Event<?>> events = getEvents();
                Set<List<? extends Event<?>>> seqsShorter = generateSequences(
                                length - 1, fromStart);
                for (Event<?> event : events) {
                        for (List<? extends Event<?>> seqShorter : seqsShorter) {
                                Event<?> lastEvent = event;
                                if (getProbability(seqShorter, lastEvent) > 0.0) {
                                        List<Event<?>> subSeq = new ArrayList<Event<?>>(seqShorter);
                                        subSeq.add(lastEvent);
                                        sequenceSet.add(subSeq);
                                }
                        }
                }
                return sequenceSet;
        }

        /*
         * (non-Javadoc)
         * 
         * @see
         * de.ugoe.cs.eventbench.models.IStochasticProcess#generateValidSequences
         * (int)
         */
        @Override
        public Set<List<? extends Event<?>>> generateValidSequences(int length) {
                // check for min-length implicitly done by generateSequences
                Set<List<? extends Event<?>>> allSequences = generateSequences(length,
                                true);
                Set<List<? extends Event<?>>> validSequences = new LinkedHashSet<List<? extends Event<?>>>();
                for (List<? extends Event<?>> sequence : allSequences) {
                        if (sequence.size() == length
                                        && Event.ENDEVENT.equals(sequence.get(sequence.size() - 1))) {
                                validSequences.add(sequence);
                        }
                }
                return validSequences;
        }

}