P-Box Theory and SVM Methods with Application in Pattern Recognition

Abstract:

In order to solve the problem of the information loss on the feature extraction process in the traditional pattern recognition, a new method based on probability boxes theory was proposed. Firstly, the skewness of the fault signal data were used as the information source to construct the tow p-boxes about X and direction. Then, to take advantage of the complementation of the information source, the tow p-boxes from different directions were fused. Finally, the SVM features database was established by extracting different types of cumulative uncertainty measures from p-boxes. The analysis result shows that the combination of p-box and SVM can achieve a high recognition rate, which makes a new way for pattern recognition.

• Periodical: Applied Mechanics and Materials (Volumes 651-653)
• Pages: 472-475
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.651-653.472
• Online since: September 2014
• Authors: Jia Man Ding, Yi Du*, Qing Xin Wang, Ying Jiang, Lian Yin Jia
• Keywords: Fusion, Pattern Recognition, Probability Box, SVM, Uncertainty

References

[1] Klir GJ: Uncertainty and Information: Foundations of Generalized Information Theory. Wiley-IEEE Press, (2005).

[2] Ferson. Constructing Probability Boxes and Dempster-Shafer Structures [M]. Sandia National Laboratories, 2003: 143-180.

DOI: 10.2172/809606

[3] Berleant, Bounding. The Times to Failure of 2-Components Systems. IEEE Transaction on Reliability, 2004, 53(4): 542-550.

DOI: 10.1109/tr.2004.837315

[4] Fulvio Tonon. Using random set theory to propagate epistemic uncertainty through a mechanical system. Reliability Engineering and System Safety, 2004, 85: 169-181.

DOI: 10.1016/j.ress.2004.03.010

[5] Kriegler, Held. Utilizing belief functions for the estimation of future climate change. International Journal of Approximate Reasoning, 2005, 39: 185-209.

DOI: 10.1016/j.ijar.2004.10.005

[6] Scott Ferson, Troy Tucker. Sensitivity in risk analyses with uncertain numbers. Applied Biomathematics, Setauket, New York, (2005).

[7] Ferson, Tucker. Sensitivity analysis using probability bounding. Reliability engineering and system safety, 2006, 91(10): 1435–1442.

DOI: 10.1016/j.ress.2005.11.052

[8] Limbourg, Savic. Fault tree analysis in an early design stage using the Dempster-Shafer theory of evidence. Risk, Reliability and Societal Safety. London: Aven & Vinnem (eds) Taylor & Francis Group, (2007).

[9] Michael Oberguggenberger, Julian King, Bernhard Schmelzer. Imprecise probability methods for sensitivity analysis in engineering. In 5th International Symposium on Imprecise Probability: Theories and Applications, Prague, Czech Republic, 2007, 6: 1130-1138.

DOI: 10.1016/j.ijar.2008.09.004

[10] Vapnik V N. Statistical learning theory [M]. Xu JianHua, Zhang XueGong, trans. Berjing: Publishing House of Electronics Industry, (2006).