Abstract | ||
---|---|---|
A new nearest-neighbor method is described for estimating the Bayes risk of a multiclass pattern claSSification problem from sample data (e.g., a classified training set). Although it is assumed that the classification prob(cid:173) lem can be accurately described by sufficiently smooth class-conditional distributions, neither these distributions, nor the corresponding prior prob(cid:173) abilities of the classes are required. Thus this method can be applied to practical problems where the underlying probabilities are not known. This method is illustrated using two different pattern recognition problems. |
Year | Venue | Field |
---|---|---|
1995 | ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 8: PROCEEDINGS OF THE 1995 CONFERENCE | Training set,Bayes' rule,Pattern recognition,Computer science,Artificial intelligence,Bayes error rate,Machine learning,Bayes' theorem |
DocType | Volume | ISSN |
Conference | 8 | 1049-5258 |
Citations | PageRank | References |
0 | 0.34 | 5 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Robert R. Snapp | 1 | 56 | 52.96 |
Tong Xu | 2 | 0 | 0.34 |