Namespace NOpenNLP.Tools.Ml.Maxent.Quasinewton
Classes
- LineSearch
Class that performs line search to find minimum
- LineSearch.LineSearchResult
Class to store lineSearch result
- NegLogLikelihood
Evaluate negative log-likelihood and its gradient from DataIndexer.
- ParallelNegLogLikelihood
Evaluate negative log-likelihood and its gradient in parallel
- QNMinimizer
Implementation of L-BFGS which supports L1-, L2-regularization and Elastic Net for solving convex optimization problems.
Usage example:
// Quadratic function f(x) = (x-1)^2 + 10 // f obtains its minimum value 10 at x = 1 IFunction f = new QuadraticFunction(); QNMinimizer minimizer = new QNMinimizer(); double[] x = minimizer.Minimize(f); double min = f.ValueAt(x);
- QNMinimizer.L2RegFunction
L2-regularized objective function
- QNTrainer
Maxent model trainer using L-BFGS algorithm.
Interfaces
- IFunction
Interface for a function
- QNMinimizer.IEvaluator
Evaluate quality of training parameters. For example, it can be used to report model's training accuracy when we train a Maximum Entropy classifier.