원문정보
초록
영어
Handwritten digit recognition is a task of great importance in many applications. There are different challenges faced while attempting to solve this problem. It has drawn much attention from the field of machine learning and pattern recognition. Minimax probability machine (MPM) is a novel method in machine learning and data mining. In this paper, we present an extension algorithm for MPM, which is named twin minimax probability machine (TWMPM). TWMPM generates two hyperplanes to improve the classification accuracy. Experiment results on several data sets from the UCI repository demonstrate that TWMPM can improve performance of MPM in most cases. The proposed method is used for recognizing the handwritten digits provided in the MNIST data set of images of handwritten digits (0-9). The testing accuracies are improved comparing with MPM.
목차
1. Introduction
2. Feature Extraction
3. TWMPM
3.1. Minimax Probability Machine (MPM)
3.2. Twin Support Vector Machine (TWSVM)
3.3. TWMPM
4. Experiments
4.1. UCI Datasets
4.2. MNIST Data Set
5. Conclusion
Acknowledgements
References