Previous |  Up |  Next


feature selection; a priori information
The paper briefly reviews recent advances in the methodology of feature selection (FS) and the conceptual base of a consulting system for solving FS problems. The reasons for designing a kind of expert or consulting system which would guide a less experienced user are outlined. The paper also attempts to provide a guideline which approach to choose with respect to the extent of a priori knowledge of the problem. The methods discussed here form the core of the software package being developed for solving FS problem. Two basic approaches are reviewed and the conditions under which they should be used are specified. One approach involves the use of the computationally effective Floating search methods. The alternative approach trades off the requirement for a priori information for the requirement of sufficient data to represent the distributions involved. Owing to its nature it is particularly suitable for cases when the underlying probability distributions are not unimodal. The approach attempts to achieve simultaneous feature selection and decision rule inference. According to the criterion adopted there are two variants allowing the selection of features either for optimal representation or discrimination.
[2] Boekee D. E., Lubbe J. C. A. Van der: Some aspects of error bounds in feature selection. Pattern Recognition 11 (1979), 252–360 DOI 10.1016/0031-3203(79)90047-5 | MR 0553356
[3] Devijver P., Kittler J.: Pattern Recognition: A Statistical Approach. Prentice 1982 MR 0692767 | Zbl 0542.68071
[4] Ferri F. J., Pudil P., Hatef M., Kittler J.: Comparative study of techniques for large–scale feature selection. In: Pattern Recognition in Practice IV (E. S. Gelsema and L. N. Kanal, eds.), Elsevier 1994, pp. 403–413
[5] Kittler J.: Feature selection and extraction. Handbook of Pattern Recognition and Image Processing (T. Y. Young and K. S. Fu, eds.), Academic Press, New York 1986, pp. 60–81
[6] Narendra P. M., Fukunaga K.: A Branch and Bound Algorithm for feature subset selection. IEEE Trans. Computers C-26 (1977), 917–922 DOI 10.1109/TC.1977.1674939
[7] Novovičová J., Pudil P., Kittler J.: Feature selection based on divergence for empirical class densities. In: Proc. of the 9th Scandinavian Conf. on Image Analysis, Uppsala 1995
[8] Novovičová J., Pudil P., Kittler J.: Divergence based feature selection for multimodal class densities. IEEE Trans. Pattern Recognition Machine Intelligence 18 (1996), 2, 218–223 DOI 10.1109/34.481557
[9] Novovičová J., Pudil P.: Feature selection and classification by modified model with latent structure. In: Dealing With Complexity: Neural Network Approach. Springer Verlag, Berlin 1997, pp. 126–140
[10] Pudil P., Bláha S., Novovičová J.: PREDITAS – software package for solving pattern recognition and diagnostic problems. In: Proc. BPRA 4th Internat. Conference on Pattern Recognition, Cambridge (J. Kittler, ed.), Springer–Verlag, Berlin 1988, pp. 146–152
[11] Pudil P., Novovičová J., Choakjarernwanit N., Kittler J.: An analysis of the Max–Min approach to feature selection. Pattern Recognition Lett. 14 (1993), 11, 841–847 DOI 10.1016/0167-8655(93)90147-6 | Zbl 0802.68118
[12] Pudil P., Novovičová J., Choakjarernwanit N., Kittler J.: The Max–Min approach to feature selection: Its foundations and potential. Indian J. Pure Appl. Math. 24 (1994), 11, 69–81
[13] Pudil P., Novovičová J., Kittler J.: Automatic machine learning of decision rule for classification problems in image analysis. In: Proceedings of BMVC ’93 – the 4th British Machine Vision Conference, 1993
[14] Pudil P., Novovičová J., Kittler J.: Simultaneous learning of decision rules and important attributes for classification problems in image analysis. Image and Vision Computing 12 (1994), 3, 193–198 DOI 10.1016/0262-8856(94)90072-8
[15] Pudil P., Ferri F., Novovičová J., Kittler J.: Floating search methods for feature selection with nonmonotonic criterion functions. In: Proc. of the 12th IAPR Intern. Conf. on Pattern Recognition, Jerusalem 1994, IEEE Comp. Society Press, pp. 279–283
[16] Pudil P., Novovičová J., Kittler J.: Floating search methods in feature selection. Pattern Recognition Lett. 15 (1994), 1119–1125 DOI 10.1016/0167-8655(94)90127-9
[17] Pudil P., Novovičová J., Choakjarerwanit N., Kittler J.: Feature selection based on the approximation of class densities by finite mixtures of special type. Pattern Recognition 28 (1995), 9, 1389–1397 DOI 10.1016/0031-3203(94)00009-B
[18] P.Pudil J.Novovičová, Ferri F. J.: Methods of dimensionality reduction in statistical pattern recognition. In: Proceedings of the IEEE European Workshop CMP’94, Prague 1994, Institute of Information Theory and Automation, pp. 185–198
[19] Siedlecki W., Sklansky J.: On automatic feature selection. Internat. J. Pattern Recognition and Artificial Intelligence 2 (1988), 2, 197–220 DOI 10.1142/S0218001488000145
[20] Zongker D., Jain A.: Algorithms for feature selection: An evaluation. In: Proceedings of 13th International Conference on Pattern Recognition, Vienna 1996, Vol. II, Track B, pp. 18–22
Partner of
EuDML logo