Difference between revisions of "Generalized Iterative Scaling"
From Cohen Courses
Jump to navigationJump to search (Created page with 'Being edited by Rui Correia') |
|||
| Line 1: | Line 1: | ||
Being edited by Rui Correia | Being edited by Rui Correia | ||
| + | |||
| + | == What is the method == | ||
| + | The Generalized Iterative scaling algorithm is a [Category:Method] | ||
| + | the circumstances under which it is meant to be used | ||
| + | |||
| + | |||
| + | you are expected to explain clearly what the method is | ||
| + | |||
| + | and list papers that use it | ||
| + | |||
| + | inputs and resources, | ||
| + | |||
| + | outputs | ||
| + | |||
| + | things the method is comparable to. | ||
| + | |||
| + | Explain what motivations or assumptions underlie the method | ||
| + | |||
| + | |||
| + | what makes it different from earlier methods. | ||
| + | |||
| + | |||
| + | |||
| + | Advantages: | ||
| + | � incorporates feature selection, | ||
| + | � scales up well in numbers of features w.r.t. other | ||
| + | methods, | ||
| + | � resilient to feature dependence. | ||
| + | Disadvantages: | ||
| + | � slow in training relative to other classi�cation | ||
| + | methods, | ||
| + | � problems with smoothing, | ||
| + | � binary and summation assumptions (not a big | ||
| + | deal) | ||
| + | |||
| + | == Related Papers == | ||
Revision as of 20:10, 29 September 2011
Being edited by Rui Correia
What is the method
The Generalized Iterative scaling algorithm is a [Category:Method] the circumstances under which it is meant to be used
you are expected to explain clearly what the method is
and list papers that use it
inputs and resources,
outputs
things the method is comparable to.
Explain what motivations or assumptions underlie the method
what makes it different from earlier methods.
Advantages: � incorporates feature selection, � scales up well in numbers of features w.r.t. other methods, � resilient to feature dependence. Disadvantages: � slow in training relative to other classi�cation methods, � problems with smoothing, � binary and summation assumptions (not a big deal)
