Semi-supervised learning with local and global consistency. (English) Zbl 1319.68174
Summary: In a lot of practical machine learning applications, such as web page classification, protein shape classification, unlabelled instances are easy to obtain, but labelled instances are rather too expensive to get. Thus, recently, semi-supervised learning (SSL) methods including graph-based algorithms have attracted many interests from researchers. However, most of these algorithms used the Gaussian function to calculate weights of the edge of the graph. In this paper, we proposed a novel weight for graph-based semi-supervised algorithms. In this new algorithm, the label information is added from problem into SSL algorithm, and the geodesic distance is utilized instead of Euclidean distance to calculate the distance between two instances. Furthermore, class prior knowledge is also added from problem into the target function. In this paper, we focus on learning with local and global consistency. We found that the effect of class prior knowledge maybe different between under low-label rate and high-label rate. Experiments on two University of California Irvine (UCI) data sets and United States Postal Service handwritten digit recognition show that our proposed algorithm is really effective.
MSC:
| 68T05 | Learning and adaptive systems in artificial intelligence |
Keywords:
machine learning; semi-supervised learning; class prior knowledge; geodesic distance; label informationSoftware:
UCI-mlReferences:
| [1] | DOI: 10.1162/089976603321780317 · Zbl 1085.68119 · doi:10.1162/089976603321780317 |
| [2] | DOI: 10.1023/B:MACH.0000033120.25363.1e · Zbl 1089.68086 · doi:10.1023/B:MACH.0000033120.25363.1e |
| [3] | Blake C., UCI Repository of Machine Learning Databases (1998) |
| [4] | DOI: 10.1109/TGRS.2006.877950 · doi:10.1109/TGRS.2006.877950 |
| [5] | DOI: 10.1109/TGRS.2007.895416 · doi:10.1109/TGRS.2007.895416 |
| [6] | Chung F.R.K., Spectral Graph Theory (1997) · Zbl 0867.05046 |
| [7] | DOI: 10.1007/11540007_1 · doi:10.1007/11540007_1 |
| [8] | DOI: 10.1109/3477.552191 · doi:10.1109/3477.552191 |
| [9] | Jaakkola M.S.T., Neural Information Processing Systems 2 pp 945– (2002) |
| [10] | DOI: 10.1109/TSMCA.2007.904579 · doi:10.1109/TSMCA.2007.904579 |
| [11] | DOI: 10.1023/A:1022627018023 · Zbl 0771.68095 · doi:10.1023/A:1022627018023 |
| [12] | DOI: 10.1109/TSMCA.2007.904745 · doi:10.1109/TSMCA.2007.904745 |
| [13] | DOI: 10.1088/1741-2560/4/3/010 · doi:10.1088/1741-2560/4/3/010 |
| [14] | DOI: 10.1109/35.41401 · doi:10.1109/35.41401 |
| [15] | DOI: 10.1109/TIP.2012.2218825 · Zbl 1373.94269 · doi:10.1109/TIP.2012.2218825 |
| [16] | Nie F., IEEE Trans. Syst. Man Cybern. Part B: Cybern 41 (99) pp 1– (2011) |
| [17] | DOI: 10.1023/A:1007692713085 · Zbl 0949.68162 · doi:10.1023/A:1007692713085 |
| [18] | DOI: 10.1126/science.290.5500.2323 · doi:10.1126/science.290.5500.2323 |
| [19] | DOI: 10.1109/36.312897 · doi:10.1109/36.312897 |
| [20] | DOI: 10.1126/science.290.5500.2319 · doi:10.1126/science.290.5500.2319 |
| [21] | Zhang X., Matrix Analysis and Applications (2004) |
| [22] | Zhou D., Neural Information Processing Systems 16 pp 321– (2004) |
| [23] | DOI: 10.1109/TKDE.2005.186 · doi:10.1109/TKDE.2005.186 |
| [24] | Zhu X., Semi-supervised learning literature survey (2005) |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
