Abstract | ||
---|---|---|
Parameter tuning through cross-validation becomes very
difficult when the validation set contains no or only a
few examples of the classes in the evaluation set. We
address this open challenge by using a combination of
classifiers with different performance characteristics to
effectively reduce the performance variance on average of
the overall system across all classes, including those not
seen before. This approach allows us to tune the
combination system on available but less-representative
validation data and obtain smaller performance degradation
of this system on the evaluation data than using a
single-method classifier alone. We tested this approach by
applying k-Nearest Neighbor, Rocchio and Language Modeling
classifiers and their combination to the event tracking
problem in the Topic Detection and Tracking (TDT) domain,
where new classes (events) are created constantly over
time, and representative validation sets for new classes
are often difficult to obtain on time. When parameters
tuned on an early benchmark TDT corpus were evaluated on a
later TDT benchmark corpus with no overlapping events, we
observed a 38-65\% reduction in tracking cost (a weighted
combination of errors) by the combined system over the
individual methods evaluated under the same conditions,
strongly suggesting the robustness of this approach as a
solution for improving cross-class performance consistency
of statistical classifiers when standard cross-validation
fails due to the lack of representative validation sets. |
Year | Venue | Keywords |
---|---|---|
2000 | ICML | combining multiple learning strategies,effective cross validation,language model,k nearest neighbor,cross validation |
Field | DocType | ISBN |
Data mining,Pattern recognition,Computer science,Artificial intelligence,Cross-validation,Machine learning | Conference | 1-55860-707-2 |
Citations | PageRank | References |
39 | 4.32 | 8 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yiming Yang | 1 | 3299 | 344.91 |
Tom Ault | 2 | 156 | 19.83 |
Thomas Pierce | 3 | 39 | 4.32 |