Feature Subset Selection Using an Optimized Hill Climbing Algorithm for Handwritten Character Recognition
This paper presents an optimized Hill Climbing algorithm to select a subset of features for handwritten character recognition. The search is conducted taking into account a random mutation strategy and the initial relevance of each feature in the recognition process. The experiments have shown a reduction in the original number of features used in an MLP-based character recognizer from 132 to 77 features (reduction of 42%) without a significant loss in terms of recognition rates, which are 99% for 60,089 samples of digits, and 93% for 11,941 uppercase characters, both handwritten samples from the NIST SD19 database. The proposed method has shown to be an interesting strategy to implement a wrapper approach without the need of complex and expensive hardware architectures.
- 2.Boz, O.: Feature Subset Selection by Using Sorted Feature Relevance. In: ICMLA International Conference on Machine Learning and Application, Las Vegas City (2002)Google Scholar
- 3.Molina, L.C., Belanche, L., Nebot, À.: Feature Selection Algorithms: Survey Experimental and Evaluation. In: IEEE International Conference on Data Mining, Maebashi City, vol. 1, pp. 306–313 (2002)Google Scholar
- 5.Aha, D.W., Bankert, R.L.: Feature Selection for Case-based Classification of Cloud Types: an empirical comparison. In: Working Notes of the AAAI, Workshop on Case-Based Reasoning, vol. 1, pp. 106–112 (1994)Google Scholar
- 6.Raman, B., Ioerger, T.R.: Enhancing Learning Using Feature and Example Selection. Journal of Machine Learning Research (2003) (submitted for publication)Google Scholar
- 8.Oliveira, L.S., Sabourin, R., Bortolozzi, F., Suen, C.Y.: Feature Selection Using Multiobjective Genetic Algorithms for Handwritten Digit Recognition. In: Proc. International Conference on Pattern Recognition, Quebec City, vol. 1, pp. 568–571 (2002)Google Scholar
- 9.Oliveira, L.S., Sabourin, R., Bortolozzi, F., Suen, C.Y.: A Methodology for Feature Selection Using Multi-objective Genetic Algorithms for Handwritten Digit String Recognition. In: The International Journal of Pattern Recognition and Artificial Intelligence (IJPRAI), vol. 17(6), pp. 903–930 (2003)Google Scholar