本文整理汇总了Java中cc.mallet.optimize.Optimizer.optimize方法的典型用法代码示例。如果您正苦于以下问题：Java Optimizer.optimize方法的具体用法？Java Optimizer.optimize怎么用？Java Optimizer.optimize使用的例子？那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类cc.mallet.optimize.Optimizer的用法示例。
在下文中一共展示了Optimizer.optimize方法的3个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Java代码示例。
示例1: train
import cc.mallet.optimize.Optimizer; //导入方法依赖的package包/类
public MCMaxEnt train (InstanceList trainingSet)
{
logger.fine ("trainingSet.size() = "+trainingSet.size());
mt = new MaximizableTrainer (trainingSet, (MCMaxEnt)initialClassifier);
Optimizer maximizer = new LimitedMemoryBFGS(mt);
// CPAL - change the tolerance for large vocab experiments
((LimitedMemoryBFGS)maximizer).setTolerance(.00001); // std is .0001;
maximizer.optimize (); // XXX given the loop below, this seems wrong.
logger.info("MCMaxEnt ngetValueCalls:"+getValueCalls()+"\nMCMaxEnt ngetValueGradientCalls:"+getValueGradientCalls());
// boolean converged;
//
// for (int i = 0; i < numIterations; i++) {
// converged = maximizer.maximize (mt, 1);
// if (converged)
// break;
// else if (evaluator != null)
// if (!evaluator.evaluate (mt.getClassifier(), converged, i, mt.getValue(),
// trainingSet, validationSet, testSet))
// break;
// }
// TestMaximizable.testValueAndGradient (mt);
progressLogger.info("\n"); // progess messages are on one line; move on.
return mt.getClassifier ();
}
示例2: forward
import cc.mallet.optimize.Optimizer; //导入方法依赖的package包/类
public void forward() {
// initialize first state
double[] mean0 = new double[dimension];
Arrays.fill(mean0, 0.0);
double[] var0 = new double[dimension];
Arrays.fill(var0, sigma);
double[] preMean = mean0;
double[] preVar = var0;
int numConverged = 0;
for (int t = 0; t < this.T; t++) {
StateObjective objective = new StateObjective(preMean, preVar, states[t].getObservations());
Optimizer optimizer = new LimitedMemoryBFGS(objective);
boolean converged = false;
try {
converged = optimizer.optimize();
} catch (Exception ex) {
// This exception may be thrown if L-BFGS
// cannot step in the current direction.
// This condition does not necessarily mean that
// the optimizer has failed, but it doesn't want
// to claim to have succeeded...
// do nothing
}
if (converged) {
numConverged++;
}
for (int i = 0; i < dimension; i++) {
states[t].setMean(i, objective.getParameter(i));
}
// compute diagonal approximation of the Hessian
double[] exps = new double[dimension];
double sumExp = 0.0;
for (int i = 0; i < dimension; i++) {
exps[i] = Math.exp(states[t].getMean(i));
sumExp += exps[i];
}
for (int i = 0; i < dimension; i++) {
double prob = exps[i] / sumExp;
double negHess =
1.0 / preVar[i]
+ states[t].getCountSum() * prob * (1 - prob);
states[t].setVariance(i, 1.0 / negHess);
}
// update
for (int i = 0; i < dimension; i++) {
preMean[i] = states[t].getMean(i);
preVar[t] = states[t].getVariance(i) + this.sigmaSquare;
}
System.out.println("State " + t + ". " + converged);
System.out.println("Mean:\t" + MiscUtils.arrayToString(states[t].getMean()));
System.out.println("Var:\t" + MiscUtils.arrayToString(states[t].getVariance()));
System.out.println("Dist:\t" + MiscUtils.arrayToString(states[t].getLogisticNormalDistribution()));
System.out.println("True:\t" + MiscUtils.arrayToString(trueDist[t]));
System.out.println("Obs:\t" + MiscUtils.arrayToString(observations[t]));
System.out.println();
}
System.out.println("# converged = " + numConverged);
}
示例3: forwardSingleChain
import cc.mallet.optimize.Optimizer; //导入方法依赖的package包/类
/**
* Perform forward filtering for a single chain of nodes
*
* @param nodes The chain
*/
private void forwardSingleChain(ArrayList<DNCRPNode> nodes,
double[] priorMean, double[] priorVar) {
if (debug) {
logln("--- forward filtering chain: " + nodes.toString());
}
// double[] preMean = zeros.clone();
// double[] preVar = sigmaSquares.clone();
double[] preMean = priorMean.clone();
double[] preVar = priorVar.clone();
int numConverged = 0;
for (int t = 0; t < nodes.size(); t++) {
StateObjective objective = new StateObjective(preMean, preVar, nodes.get(t).getContent().getSparseCounts());
Optimizer optimizer = new LimitedMemoryBFGS(objective);
boolean converged = false;
try {
converged = optimizer.optimize();
} catch (Exception ex) {
// This exception may be thrown if L-BFGS
// cannot step in the current direction.
// This condition does not necessarily mean that
// the optimizer has failed, but it doesn't want
// to claim to have succeeded...
// do nothing
}
if (converged) {
numConverged++;
}
for (int i = 0; i < V; i++) {
nodes.get(t).getContent().setMean(i, objective.getParameter(i));
}
// compute diagonal approximation of the Hessian
double[] exps = new double[V];
double sumExp = 0.0;
for (int i = 0; i < V; i++) {
exps[i] = Math.exp(nodes.get(t).getContent().getMean(i));
sumExp += exps[i];
}
for (int i = 0; i < V; i++) {
double prob = exps[i] / sumExp;
double negHess =
1.0 / preVar[i]
+ nodes.get(t).getContent().getCountSum() * prob * (1 - prob);
nodes.get(t).getContent().setVariance(i, 1.0 / negHess);
// logln("i = " + i
// + ". exps = " + MiscUtils.formatDouble(exps[i])
// + ". preVar = " + MiscUtils.formatDouble(preVar[i])
// + ". prob = " + MiscUtils.formatDouble(prob)
// + ". negH = " + MiscUtils.formatDouble(negHess)
// + ". ---> " + MiscUtils.formatDouble(1.0 / negHess));
}
// debug
// logln("---> node: " + nodes.get(t).toString());
// logln(MiscUtils.arrayToString(nodes.get(t).getContent().getVariance()) + "\n");
// update
nodes.get(t).getContent().updateDistribution();
for (int i = 0; i < V; i++) {
preMean[i] = nodes.get(t).getContent().getMean(i);
preVar[i] = nodes.get(t).getContent().getVariance(i) + sigmaSquares[0];
}
}
}