Inheritance diagram for classias::train::lbfgs_logistic_multi< data_tmpl, model_tmpl >:
| data_tmpl | The type of the data set for training. | |
| model_tmpl | The type of the feature weights. |
Public Member Functions | |
| lbfgs_logistic_multi () | |
| Constructs the object. | |
| virtual | ~lbfgs_logistic_multi () |
| Destructs the object. | |
| void | clear () |
| Resets the internal states and parameters to default. | |
| void | train (const data_type &data, std::ostream &os, int holdout=-1, bool acconly=true) |
| Trains a model on a data set. | |
Protected Types | |
| typedef data_tmpl | data_type |
| A type representing a data set for training. | |
| typedef model_tmpl | model_type |
| The type implementing a model (weight vector for features). | |
| typedef lbfgs_base< model_tmpl > | base_class |
| A synonym of the base class. | |
|
typedef lbfgs_logistic_multi< data_type, model_tmpl > | this_class |
| A synonym of this class. | |
| typedef model_type::value_type | value_type |
| The type representing a value. | |
| typedef data_type::instance_type | instance_type |
| A type representing an instance in the training data. | |
| typedef data_type::const_iterator | const_iterator |
| A type providing a read-only random-access iterator for instances. | |
| typedef instance_type::attributes_type | attributes_type |
| A type representing an attribute. | |
| typedef data_type::feature_generator_type | feature_generator_type |
| A type representing a feature generator. | |
| typedef data_type::attribute_type | attribute_type |
| A type representing a candidate for an instance. | |
|
typedef classify::linear_multi_logistic< model_type > | error_type |
| The type of a classifier. | |
Protected Member Functions | |
| virtual value_type | loss_and_gradient (const value_type *x, value_type *g, const int n) |
| Computes the loss and gradients of the data set. | |
| void | holdout_evaluation () |
| Performs a holdout evaluation. | |
| template<class feature_generator_type, class iterator_type> | |
| void | add_weights (value_type *w, int l, feature_generator_type &fgen, iterator_type first, iterator_type last, value_type delta) |
| Adds a value to weights associated with a feature vector. | |
Protected Attributes | |
| value_type * | m_oexps |
| An array [K] of observation expectations. | |
| const data_type * | m_data |
| A data set for training. | |
| bool | m_acconly |
| The flag indicating whether. | |
| virtual value_type classias::train::lbfgs_logistic_multi< data_tmpl, model_tmpl >::loss_and_gradient | ( | const value_type * | x, | |
| value_type * | g, | |||
| const int | n | |||
| ) | [inline, protected, virtual] |
Computes the loss and gradients of the data set.
| x | The current feature weights. | |
| g | The gradient vector to which this function stores. | |
| n | The number of features. |
Implements classias::train::lbfgs_base< model_tmpl >.
| void classias::train::lbfgs_logistic_multi< data_tmpl, model_tmpl >::train | ( | const data_type & | data, | |
| std::ostream & | os, | |||
| int | holdout = -1, |
|||
| bool | acconly = true | |||
| ) | [inline] |
Trains a model on a data set.
| data | The data set for training (and holdout evaluation). | |
| os | The output stream for progress reports. | |
| holdout | The group number for holdout evaluation. Specify a negative value if a holdout evaluation is unnecessary. | |
| acconly | The flag indicating whether precision, recall, and F1 scores are unnecessary. |
| void classias::train::lbfgs_logistic_multi< data_tmpl, model_tmpl >::add_weights | ( | value_type * | w, | |
| int | l, | |||
| feature_generator_type & | fgen, | |||
| iterator_type | first, | |||
| iterator_type | last, | |||
| value_type | delta | |||
| ) | [inline, protected] |
Adds a value to weights associated with a feature vector.
| w | The weight vector to which an update occurs. | |
| l | The candidate index. | |
| fgen | The feature generator. | |
| first | The iterator pointing to the first element of the feature vector. | |
| last | The iterator pointing just beyond the last element of the feature vector. | |
| delta | The value to be added to the weights. |