A New Framework of Smoothed Location Model with Multiple Correspondence Analysis
Theimplication of a considering large binary variabIes into the smoat€td location model wiU create too many multinomid &Us or lead to'high mnltinomid cells and mom worrying is that it will cause most of them are empty. We refer this sitnation as large sparsity pmblem. When large sparsity...
| Main Author: | |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | https://repo.uum.edu.my/id/eprint/24453/1/iCMS%202015%20117-127.pdf |
| Summary: | Theimplication of a considering large binary variabIes into the smoat€td
location model wiU create too many multinomid &Us or lead to'high mnltinomid
cells and mom worrying is that it will cause most of them are empty. We refer this
sitnation as large sparsity pmblem. When large sparsity of midhomial d s occurs,
the smoothed estimators of location model will be greatly biased, hence creating
ffmtmting performance. At worsC the classscation rules m o t be constructed.
This issue has attracted this paper to fWAer investigate and propose a new approach
of the smoothed location model when facing with largk sparsiw prob1em. |
|---|