NextCloud/dlib
1
0
Fork 0
mirror of https://github.com/davisking/dlib.git synced 2024-11-01 10:14:53 +08:00
Code Issues Packages Projects Releases Wiki Activity

Clarified spec

Browse Source
This commit is contained in:
Davis King 2012-03-11 10:37:21 -04:00
parent 4c93d6765a
commit 6ab5882047
1 changed files with 9 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
dlib/filtering/rls_filter_abstract.h
View File

@ -26,7 +26,7 @@ namespace dlib
determine how to best combine the contents of the history buffer to determine how to best combine the contents of the history buffer to
predict each point. Therefore, each time update() is called with predict each point. Therefore, each time update() is called with
a point, recursive least squares updates the linear combination weights, a point, recursive least squares updates the linear combination weights,
and then we insert the point into the history buffer. After that, the and then it inserts the point into the history buffer. After that, the
next prediction is based on these updated weights and the current history next prediction is based on these updated weights and the current history
buffer. buffer.
!*/ !*/
@ -38,8 +38,8 @@ namespace dlib
/*! /*!
ensures ensures
- #get_window_size() == 5 - #get_window_size() == 5
- #get_c() == 100
- #get_forget_factor() == 0.8 - #get_forget_factor() == 0.8
- #get_c() == 100
- #get_predicted_next_state().size() == 0 - #get_predicted_next_state().size() == 0
!*/ !*/
@ -84,11 +84,11 @@ namespace dlib
linearly combining the history buffer into a prediction of the next point. linearly combining the history buffer into a prediction of the next point.
- else - else
- Old calls to update(z) are eventually forgotten. That is, the smaller - Old calls to update(z) are eventually forgotten. That is, the smaller
the forget factor, the less the recursive least squares algorithm will the forget factor, the less recursive least squares will care about
care about attempting to find linear combination weights which would have attempting to find linear combination weights which would have make
make good predictions on old points. It will care more about fitting good predictions on old points. It will care more about fitting recent
recent points. This is appropriate if the statistical properties of points. This is appropriate if the statistical properties of the time
the time series we are modeling are not constant. series we are modeling are not constant.
!*/ !*/
unsigned long get_window_size ( unsigned long get_window_size (
@ -104,7 +104,7 @@ namespace dlib
/*! /*!
ensures ensures
- Propagates the prediction forward in time. - Propagates the prediction forward in time.
- In particular, the value in #get_predicted_next_state() is inserted - In particular, the value in get_predicted_next_state() is inserted
into the history buffer and then the next prediction is estimated into the history buffer and then the next prediction is estimated
based on this updated history buffer. based on this updated history buffer.
- #get_predicted_next_state() == the prediction for the next point - #get_predicted_next_state() == the prediction for the next point
@ -130,6 +130,7 @@ namespace dlib
these updated weights and history buffer. these updated weights and history buffer.
- #get_predicted_next_state() == the prediction for the next point - #get_predicted_next_state() == the prediction for the next point
in the time series. in the time series.
- #get_predicted_next_state().size() == z.size()
!*/ !*/
const matrix<double,0,1>& get_predicted_next_state( const matrix<double,0,1>& get_predicted_next_state(