![]() |
Data Mining and Machine Learning Program 2003-04 |
| Group Leader: Bertrand Clarke | Meetings: Wednesdays at 11:00am@SAMSI |
| General description of aims and activities |
|
Short Fat Hairy Data refers to the setting in which we have too
many models that are too big and too few data that are too small.
Serious people refer to this situation as high dimension low sample
size, or large p small n. Whatever you call it, it is an emerging
field of recognized importance and already a lot of work has been done
in this general area.
As a service to the community, we hope to: 1) compile a matrix of the most important techniques and measures of performance used with SFH data; 2) For each cell in the matrix, do a literature review and possibly computations; 3) Summarize the results in a catalog to help guide users. In addition to this, we recognize the importance of providing a conceptual unity to an emerging field to parallel the compendium of methods and their performance. Consequently, we hope to provide a general theoretical framework for these methods by an examination of model uncertainty. Please your comments, suggestions and general ideas to Bertrand. For anything else related to this group, contact Ernest. |
| What is happening? |
|
| Archive of Public Resources | Internal News and Resources |
| Emerging Clusters |
| Literature review | The Matrix | The Catalog | Model Uncertainty |
|
Leader: ???? Members:
|
Leader: ???? Members:
|
Leader: ???? Members:
|
Leader: Bertrand Clarke Members:
|