Name
Abstract | ||
|---|---|---|
In this paper, we present a novel semi-supervised smooth harmonic transductive learning algorithm that can get closed-form solution. Our method introduces the unlabeled class information to the learning process and tries to exploit the similar configurations shared by the label distribution of data. After discovering the property of smooth harmonic function based on spectral clustering in classification task, we design an adaptive thresholding method for smooth harmonic transductive learning based on classification error. The proposed adaptive thresholding method can select the most suitable thresholds flexibly. Plentiful experiments on data sets show our proposed closedform smooth harmonic transductive learning framework get excellent improvement compared with two baseline methods. |
Year | DOI | Venue |
|---|---|---|
2013 | 10.4304/jcp.8.12.3079-3085 | JOURNAL OF COMPUTERS |
Keywords | Field | DocType |
harmonic function, transductive learning, adaptive threshold | Transduction (machine learning),Harmonic function,Data set,Semi-supervised learning,Pattern recognition,Computer science,Harmonic,Exploit,Artificial intelligence,Machine learning | Journal |
Volume | Issue | ISSN |
8 | 12 | 1796-203X |
Citations | PageRank | References |
2 | 0.38 | 12 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ying Xie | 1 | 47 | 14.48 |
| Bin Luo | 2 | 802 | 107.57 |
| Rongbin Xu | 3 | 37 | 10.01 |
| Sibao Chen | 4 | 127 | 13.42 |