VwAdaptiveLearner uses an adaptive subgradient technique to update weights.
It uses two elements in the weight vector per feature and maintains individual learning rates for each feature. For details, refer to the VW tutorial.
Definition at line 31 of file VwAdaptiveLearner.h.
Public Member Functions | |
CVwAdaptiveLearner () | |
CVwAdaptiveLearner (CVwRegressor *regressor, CVwEnvironment *vw_env) | |
virtual | ~CVwAdaptiveLearner () |
virtual void | train (VwExample *&ex, float32_t update) |
virtual const char * | get_name () const |
Default constructor
Definition at line 20 of file VwAdaptiveLearner.cpp.
CVwAdaptiveLearner | ( | CVwRegressor * | regressor, | |
CVwEnvironment * | vw_env | |||
) |
Constructor, initializes regressor and environment
regressor | regressor to use | |
vw_env | environment to use |
Definition at line 25 of file VwAdaptiveLearner.cpp.
~CVwAdaptiveLearner | ( | ) | [virtual] |
Destructor
Definition at line 30 of file VwAdaptiveLearner.cpp.
virtual const char* get_name | ( | void | ) | const [virtual] |
Return the name of the object
Reimplemented from CVwLearner.
Definition at line 65 of file VwAdaptiveLearner.h.
Train on one example, given the update
ex | example | |
update | the update |
Implements CVwLearner.
Definition at line 34 of file VwAdaptiveLearner.cpp.