Class LDA implements regularized Linear Discriminant Analysis.
LDA learns a linear classifier and requires examples to be CSimpleFeatures. The learned linear classification rule is optimal under the assumption that both classes a gaussian distributed with equal co-variance. To find a linear separation 
in training, the in-between class variance is maximized and the within class variance is minimized, i.e.
![\[ J({\bf w})=\frac{{\bf w^T} S_B {\bf w}}{{\bf w^T} S_W {\bf w}} \]](form_5.png)
is maximized, where
![\[S_b := ({\bf m_{+1}} - {\bf m_{-1}})({\bf m_{+1}} - {\bf m_{-1}})^T \]](form_6.png)
is the between class scatter matrix and
![\[S_w := \sum_{c\in\{-1,+1\}}\sum_{{\bf x}\in X_{c}}({\bf x} - {\bf m_c})({\bf x} - {\bf m_c})^T \]](form_7.png)
is the within class scatter matrix with mean 
and 
the set of examples of class c.
LDA is very fast for low-dimensional samples. The regularization parameter 
(especially useful in the low sample case) should be tuned in cross-validation.
Definition at line 52 of file LDA.h.
Public Member Functions | |
| CLDA (float64_t gamma=0) | |
| CLDA (float64_t gamma, CSimpleFeatures< float64_t > *traindat, CLabels *trainlab) | |
| virtual | ~CLDA () |
| void | set_gamma (float64_t gamma) |
| float64_t | get_gamma () |
| virtual EClassifierType | get_classifier_type () |
| virtual void | set_features (CDotFeatures *feat) |
| virtual const char * | get_name () const |
Protected Member Functions | |
| virtual bool | train_machine (CFeatures *data=NULL) |
Protected Attributes | |
| float64_t | m_gamma |
| CLDA | ( | float64_t | gamma, | |
| CSimpleFeatures< float64_t > * | traindat, | |||
| CLabels * | trainlab | |||
| ) |
| virtual EClassifierType get_classifier_type | ( | ) | [virtual] |
| virtual const char* get_name | ( | void | ) | const [virtual] |
| virtual void set_features | ( | CDotFeatures * | feat | ) | [virtual] |
| void set_gamma | ( | float64_t | gamma | ) |
| bool train_machine | ( | CFeatures * | data = NULL |
) | [protected, virtual] |
