Name
Abstract | ||
|---|---|---|
Meta-learning involves the construction of a classifler that predicts the performance of another classifler. Previously proposed ap- proaches do this by making a single prediction (such as the expected accuracy) for a complete data set. We suggest modifying this framework so that the meta-classifler predicts for each data point in the data set whether a particular base-classifler will classify it correctly or not. While this information can be converted into a standard meta-learning out- put such as an overall accuracy estimate for the complete data set, the approach has the added advantage of providing more flne-grained infor- mation which promises to be useful in Multiple Classifler Selection and Semi-Supervised Learning. This paper describes the new framework and reports the results of an initial evaluation on a medium-sized database of classiflcation data sets. |
Year | Venue | Keywords |
|---|---|---|
2007 | MLDM Posters | semi supervised learning |
Field | DocType | Citations |
Data set,Computer science,Artificial intelligence,Classifier (linguistics),Machine learning | Conference | 0 |
PageRank | References | Authors |
0.34 | 5 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Irene M. Cramer | 1 | 14 | 2.91 |
| Barbara Rauch | 2 | 6 | 1.24 |
| Hagen Fürstenau | 3 | 533 | 20.43 |
| Dan Shen | 4 | 0 | 0.34 |
| Maria Staudte | 5 | 54 | 7.37 |