Feature Selection Using Histogram-Based Multi-objective GA for Handwritten Devanagari Numeral Recognition

摘要

In this paper, we propose an efficient feature selection method, called Histogram-Based Multi-objective Genetic Algorithm (HMOGA), for finding informative features from high-dimensional data which also improves the classification accuracy. This approach is applied on two previously proposed feature sets for handwritten Devanagari numeral recognition problem. With the feature set selected by HMOGA, final recognition is performed using the Multi-layer Per-ceptron (MLP)-based classifier. The rise in classification accuracy using only 50% of the original feature vector portrays the applicability of the developed idea for multi-objective optimization.