Abstract | ||
---|---|---|
A novel solution to the problem whether to transfer based on exchangeablity test.Statistically testing if the source data is generated from the target distribution.The test is non-parametric and distribution free.Empirically justified the proposed test is effective for predicting transfer result. This paper introduces a non-parametric test to decide whether to transfer data from a source domain to a target domain to improve the generalization performance of predictive models on the target domain. The test is based on the conformal prediction framework: it statistically tests whether the target and source data are generated from the same distribution under the exchangeability assumption. The experiments show that the test is capable of outperforming existing methods when it decides on instance transfer. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1016/j.patrec.2016.12.021 | Pattern Recognition Letters |
Keywords | Field | DocType |
Instance-transfer learning,Conformity prediction framework,Exchangeability test | Data mining,Source data,Conformal map,Artificial intelligence,Mathematics,Machine learning | Journal |
Volume | Issue | ISSN |
88 | C | 0167-8655 |
Citations | PageRank | References |
1 | 0.37 | 15 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shuang Zhou | 1 | 10 | 5.02 |
Evgueni N. Smirnov | 2 | 24 | 20.38 |
Gijs Schoenmakers | 3 | 41 | 7.21 |
Kurt Driessens | 4 | 489 | 34.75 |
Ralf L. M. Peeters | 5 | 62 | 22.61 |