한글데이타의 정보량 분석에 관한 연구

Abstract

In this thesis, the informations are computed by m-th order Markov Model with six kinds of sample data made from 2 byte combinational Hangeul to analyze the informations of Hangeul Data. The methods of analysis and results are as follows: firstly, several probabilities on the syllables of Hangeul, several probabilities, joint probabilities, conditional probabilities on the elements of syllables and conditional probabilities on the elements of inter-syllables are computed by computation program. From the computed probabilities entropy, joint entropy and conditional entropy on the elements of syllables are computed. Conditional entropy on the elements of inter-syllables and the elements of syllables including special character are also computed. Finally, average mutual information are computed for two variables and between a random variables and other two random variables. In the each conditional entropy, the highest values are appeared at conditional entropy of ‘Choseong’, conditional entropy of ‘Choseong-Jungseong’ and entropy of ‘Choseong’ respectively when the ‘Jongseong’ and ‘Jungseong-Jongseong’ are given. Conditional entropy of next syllable ‘Jungseong’ in the conditional entropy between high when the ‘Jungseong’ is given syllable. The conditional entropy between ‘Jongseong’ and ‘Choseong’ including special character is lower than that between ‘Jongseong’ and ‘Choseong’ except special character. The average mutual information between two random variables is large for ‘Choseong’ and ‘Jongseong’ and small for ‘Jongseong’ and next syllable ‘Choseong’. Average mutual information between a probability variables and other two probability variables is low at ‘Choseong-Jungseong’.