A Spheriform Quantization Method Based on Sub-Region Inherent Dimension

Abstract:

Quantization methods are very significant mining task for presenting some operations, i.e., learning and classification in machine learning and data mining since many mining and learning methods in such fields require that the testing data set must include the partitioned features. In this paper, we propose a spheriform quantization method based on sub-region inherent dimension, which induces the quantified interval number and size in data-driven way. The method assumes that a quantified cluster of points can be contained in a lower intrinsic m-dimensional spheriform space of expected radius. These sample points in the spheriform can be obtained by adaptively selecting the neighborhood at initial observation based on sub-region inherent dimension. Experimental results and analysis on UCI real data sets demonstrate that our method significantly enhances the accuracy of classification than traditional quantization methods by implementing C4.5 decision tree.

• Periodical: Applied Mechanics and Materials (Volumes 556-562)
• Pages: 4244-4247
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.556-562.4244
• Online since: May 2014
• Authors: Zhen Duo Wang*, Jing Wang, Huan Wang
• Keywords: Decision Tree, Machine Learning, Quantization, Sub-Region Inherent Dimension

References

[1] J. R. Quinlan, C4. 5: Programs for Machine Learning. San Mateo, Calif.: Morgan Kaufmann, (1993).

[2] J. Dougherty, R. Kohavi, and M. Sahami, Supervised and unsupervised discretization of continuous feature, Machine learning: Proc. 12th Intl Conf., p.194–202, (1995).

DOI: 10.1016/b978-1-55860-377-6.50032-3

[3] C. T. Su and J. H. Hsu, An extended chi2 algorithm for discretization of real value attributes, IEEE Transactions on Knowledge and Data Engineering, vol. 17, no. 3, p.437–441, (2005).

DOI: 10.1109/tkde.2005.39

[4] C. J. Tsai, C. I. Lee, and W. P. Yang, A discretization algorithm based on class-attribute contingency coefficient, Information Sciences, vol. 178, no. 19, p.714–731, (2008).

DOI: 10.1016/j.ins.2007.09.004

[5] U. Fayyad and K. Irani, Multi-interval discretization of continuous-valued attributes for classification learning, In Proc. Thirteenth International Joint Conference on Artificial Intelligence. San Mateo, CA: Morgan Kaufmann, p.1022–1027, (1993).

[6] I. T. Jolliffe, Principal component analysis, Springer-Verlag, New York, (1986).

[7] E. Levina and P. J. Bickel, Maximum likelihood estimation of intrinsic dimension, Advances in NIPS, vol. 17, (2005).