Java DoubleMatrix.putRow方法代码示例

import org.jblas.DoubleMatrix; //导入方法依赖的package包/类
/**
 * Return a numerical table with numBlobs clusters, every cluster containing numPointsPerBlob n-dimensional
 * points sampled from a Gaussian centered at a random point in the nD unit hyperbox, with a standard deviation
 * of stdDev.
 */
public static ITable getNdGaussianBlobs(int numBlobs, int numPointsPerBlob, int n, double stdDev) {
    DoubleMatrix data = new DoubleMatrix(numBlobs * numPointsPerBlob, n);
    Random rnd = new Random(42);
    for (int i = 0; i < numBlobs; i++) {
        DoubleMatrix center = new DoubleMatrix(1, n);
        for (int j = 0; j < n; j++) {
            center.put(j, 20 * rnd.nextDouble());
        }
        for (int j = 0; j < numPointsPerBlob; j++) {
            DoubleMatrix delta = new DoubleMatrix(1, n);
            for (int k = 0; k < n; k++) {
                delta.put(k, rnd.nextGaussian() * stdDev);
            }
            data.putRow(i * numPointsPerBlob + j, center.add(delta));
        }
    }
    return BlasConversions.toTable(data);
}
 

import org.jblas.DoubleMatrix; //导入方法依赖的package包/类
private static DoubleMatrix loadMatrix(String dir, int dim) throws FileNotFoundException {
	
	String input = "etc/"+dir+"/D"+dim+".csv";
	
	Scanner in = new Scanner(new File(input));
	int length = 0;
	while(in.hasNextLine()) {
		in.nextLine();
		length++;
	}
	System.out.println(length + " x "+dim);
	DoubleMatrix t = new DoubleMatrix(length, dim);
	in.close();
	in = new Scanner(new File(input));
	for(int i=0; in.hasNextLine(); i++) {
		String[] coord = in.nextLine().split("\t");
		DoubleMatrix v = new DoubleMatrix(1, coord.length);
		for(int j=0; j<coord.length; j++)
			v.put(j, Double.parseDouble(coord[j]));
		t.putRow(i, v);
	}
	in.close();

	return t;
}
 

import org.jblas.DoubleMatrix; //导入方法依赖的package包/类
public void learnWhiteningMatrix(ArrayList<T> data) {
	this.means = getMeans(data);
	DoubleMatrix X = new DoubleMatrix(data.size(), this.dimension);
	// Center observations w.r.t. their means
	int row = 0;
	for(T t : data) {
		DoubleMatrix x_i = this.getOutput(t);
		X.putRow(row++, x_i.subi(means, x_i).transpose());
	}
	// Sphericize data
	DoubleMatrix M = X.transpose().mmul(X);
	M.divi(data.size(), M);
	DoubleMatrix[] ED = Eigen.symmetricEigenvectors(M);
	DoubleMatrix E = ED[0];
	DoubleMatrix D = ED[1];

	//DoubleMatrix D_invrt = Solve.pinv(sqrt(D));
	DoubleMatrix D_invrt = diagPow(D, -0.5);
	
	this.whitener = E.mmul(D_invrt);
	this.wasWhitened = true;
	
	checkSampleCovariance(X.mmul(this.whitener));
	
	//DoubleMatrix X_std = X.mmul(E).mmul(D_invrt);
}
 

import org.jblas.DoubleMatrix; //导入方法依赖的package包/类
public static StringEmbeddings getStringEmbeddingsByPCA(StringEmbeddings embeddings, int dimension) {
	ArrayList<String> keys = new ArrayList<String>();
	HashMap<String, DoubleMatrix> original = embeddings.getAllVectors();
	keys.addAll(original.keySet());
	if(keys.size() == 0 || original.get(keys.get(0)).length == dimension) {
		return embeddings;
	}
	int originalDimension = original.get(keys.get(0)).length;
	DoubleMatrix mat = new DoubleMatrix(keys.size(), originalDimension);
	for(int r=0; r<keys.size(); r++){
		mat.putRow(r, original.get(keys.get(r)).transpose());
	}
	DoubleMatrix reduced = reduceByPCA(mat, dimension);
	StringEmbeddings newEmbeddings = new StringEmbeddings();
	for(int r=0; r<keys.size(); r++){
		newEmbeddings.addEmbedding(keys.get(r), reduced.getRow(r).transpose());
	}
	return newEmbeddings;
}
 

import org.jblas.DoubleMatrix; //导入方法依赖的package包/类
public DoubleMatrix createVectorMatrix(String[] words, WordVectors v) {
	DoubleMatrix m = new DoubleMatrix(words.length, columns);
	for (int i = 0; i < words.length; i ++) {
		m.putRow(i, new DoubleMatrix(v.getWordVector(words[i])));
	}
	return m;
}