Abstract | ||
---|---|---|
Creating an effective ensemble of clauses for large, skewed data sets requires finding a diverse, high-scoring set of clauses and then combining them in such a way as to maximize predictive performance. We have adapted the RankBoost algorithm in order to maximize area under the recall-precision curve, a much better metric when working with highly skewed data sets than ROC curves. We have also explored a range of possibilities for the weak hypotheses used by our modified RankBoost algorithm beyond using individual clauses. We provide results on four large, skewed data sets showing that our modified RankBoost algorithm outperforms the original on area under the recall-precision curves. |
Year | Venue | Keywords |
---|---|---|
2009 | ILP | recall-precision curve,individual clause,modified rankboost algorithm,weak hypothesis,skewed data set,skewed data,roc curve,effective ensemble,rankboost algorithm,predictive performance,first-order clause,learning to rank,boosting,first order |
Field | DocType | Volume |
Learning to rank,Data mining,Data set,Receiver operating characteristic,Computer science,First order,Boosting (machine learning),Artificial intelligence,Machine learning | Conference | 5989 |
ISSN | ISBN | Citations |
0302-9743 | 3-642-13839-X | 0 |
PageRank | References | Authors |
0.34 | 8 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Louis Oliphant | 1 | 43 | 3.44 |
Elizabeth S. Burnside | 2 | 199 | 27.84 |
Jude W. Shavlik | 3 | 3057 | 619.89 |