利用虛擬樣本與線性獨立解決小樣本問題

Using virtual sample and linear independence to solve small sample size problem

羅兆為(Chao-Wei Luo)

指導教授 ： 姚志佳

朝陽科技大學/資訊學院/資訊工程系/碩士(2015年)

摘要

本論文研究方向以小樣本問題為基礎，小樣本問題之所以難以處理，原因在於樣本與樣本之間存在著顯著的資料間距，同時也由於樣本小，導致無法利用統計學的方法去估算母體的分配，故以往應用於大樣本的方法不適用於小樣本上。在樣本數不足的部分，本論文將藉由生成虛擬樣本，以此獲得足夠的樣本數以及填補樣本與樣本之間的間距，再透過對樣本中各類別資料與群中心距離的計算來篩選雜訊資料，以及藉由支撐向量之特性來篩選資料，篩選完後再運用線性獨立選取支撐向量，並在實驗結果證明本研究的方法勝過其他方法。

關鍵字

線性獨立 ； 支撐向量機器 ； 小樣本 ； 虛擬樣本

並列摘要

This research proposed a novel algorithm to solve small sample size problem. The small sample size problem is difficult to solve due to it can’t use statistical methods to estimate the distribution of training samples. Therefore, the conventional method which applies to the large sample problem does not apply to the small sample size problem. This research generates virtual samples to increase the number of samples, and then calculating the probability of noise in order to filter data. After filtering, using linearly independence to select support vector. The experimental results indicate that the proposed method is effective.

並列關鍵字

Linear independent ； Support Vector Machines ； Small sample size ； Virtual sample

參考文獻

[1] B. Schölkopf and A. J. Smola, Learning with kernels: Support vector machines, regularization,optimization, and beyond, Cambridge, Mass: MIT Press, London, 2002.

[2] V. N. Vapnik, Statistical learning theory, Wiley, New York, 1998.

[3] J. Liu and S. Chen, “Discriminant common vectors versus neighborhood components analysis and laplacianfaces: A comparative study in small sample size problem,” Image and Vision Computing, vol. 24, 2006, pp. 249-262.

[4] J. Wang and K. N. Plataniotis, J. Lu and A. N. Venetsanopoulos, “Kernel quadratic discriminant analysis for small sample size problem,” Pattern Recognition, vol. 41, 2008, pp. 1528-1538.

[5] B. C. Kuo and D. A. Landgrebe, “Nonparametric weighted feature extraction for classification,” IEEE. Trans. on Geoscience and Remote Sensing, vol. 42, no. 5, 2004, pp. 1096-1105.