Examples of edu.ucla.sspace.similarity.SimilarityFunction

edu.ucla.sspace.similarity.SimilarityFunction
An Interface for any similarity metric between two {@link Vectors}. Methods should provide efficient support for sparse vectors if an implementation is feasible as these metrics are heavily used in some algorithms. Metrics should note via {@link #isSymmetric} whether or not they aresymmetic, whereby {@code method.sim(A,B) == method.sim(B,A)} alwaysholds true for any {@code A,B}. If the metric takes in any parameters, such as weights, they can use the {@link #setParams} method, which can takein any number of {@code double} arguments. Methods should also provide a noargument constructor that has reasonable default values, or loads values from system properties. @author Keith Stevens

        AffinityMatrixCreator creator = ReflectionUtil.getObjectInstance(
                argOptions.getStringOption('e'));
        if (argOptions.hasOption("edgeTypeParam"))
            creator.setParams(argOptions.getIntOption("edgeTypeParam"));


        SimilarityFunction edgeSim = new CosineSimilarity();
        SimilarityFunction kernelSim = ReflectionUtil.getObjectInstance(
                argOptions.getStringOption("edgeWeighting"));
        if (argOptions.hasOption("edgeWeighting"))
            kernelSim.setParams(argOptions.getIntOption("edgeWeightingParam"));
        creator.setFunctions(edgeSim, kernelSim);


        return new LocalityPreservingCooccurrenceSpace(creator);
    }

View Full Code Here

        this.creator = creator;
        this.linkCluster = new LinkClustering(); 
    }


    public static AffinityMatrixCreator createDefaultAffinityMatrixCreator() {
        SimilarityFunction simFunc = new CosineSimilarity();
        AffinityMatrixCreator creator =
            new NearestNeighborAffinityMatrixCreator();
        creator.setParams(10);
        creator.setFunctions(simFunc, simFunc);
        return creator;

View Full Code Here

            } catch (NumberFormatException nfe) {
                throw new IllegalArgumentException("Invalid beta", nfe);
            }
        }


        SimilarityFunction simFunc = DEFAULT_SIMILARITY_FUNCTION;
        String simFuncProp = props.getProperty(SIMILARITY_FUNCTION_PROPERTY);
        if (simFuncProp != null) {
            try {
                SimilarityFunction sf = ReflectionUtil
                    .<SimilarityFunction>getObjectInstance(simFuncProp);
                simFunc = sf;
            } catch (Error e) {
                throw new IllegalArgumentException(
                    "Invalid similarity function class", e);

View Full Code Here

        AffinityMatrixCreator creator = ReflectionUtil.getObjectInstance(
                argOptions.getStringOption('e'));
        if (argOptions.hasOption("edgeTypeParam"))
            creator.setParams(argOptions.getIntOption("edgeTypeParam"));


        SimilarityFunction edgeSim = new CosineSimilarity();
        SimilarityFunction kernelSim = ReflectionUtil.getObjectInstance(
                argOptions.getStringOption("edgeWeighting"));
        if (argOptions.hasOption("edgeWeighting"))
            kernelSim.setParams(argOptions.getIntOption("edgeWeightingParam"));
        creator.setFunctions(edgeSim, kernelSim);


        try {
            return new LocalityPreservingSemanticAnalysis(creator);
        } catch (IOException ioe) {

View Full Code Here

TOP

Related Classes of edu.ucla.sspace.similarity.SimilarityFunction

edu.ucla.sspace.clustering.DataMatrixLinkClustering

edu.ucla.sspace.clustering.FastStreamingKMeans

edu.ucla.sspace.mains.LpcsMain

edu.ucla.sspace.mains.LpsaMain

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.