논문 국내 국내전문학술지(KCI급) DERIVING A NEW DIVERGENCE MEASURE FROM EXTENDED CROSS-ENTROPY ERROR FUNCTION
- 학술지 구분 국내전문학술지(KCI급)
- 게재년월 2015-06
- 저자명 오상훈, HIROSHI WAKUYA, 박선규, 노황우, 유재수, 민병원, 오용선
- 학술지명 INTERNATIONAL JOURNAL OF CONTENTS
- 발행처명 한국콘텐츠학회
- 발행국가 국내
- 논문언어 한국어
홈페이지논문 초록 (Abstract)
Relative entropy is a divergence measure between two probability density functions of a random variable. Assuming that therandom variable has only two alphabets, the relative entropy becomes a cross-entropy error function that can accelerate training conve rgence of multi-layer perceptron neural networks. Also, the n-th order extension of cross-entropy (nCE) error function exhibits an imp roved performance in viewpoints of learning convergence and generaliz ation capability. In this paper, we derive a new divergence measure b etween two probability density functions from the nCE error function. And the new divergence measure is compared with the relative entropy t hrough the use of three-dimensional plots.