Abstract
In this paper, the occurring process of important terms in junior-high and high school science textbooks has been analyzed. I calculated total information of text by calculating the difference between mutual information of term and paragraph and that computed on the text obtained by shuffling all term positions. I regard the degree of contribution of each term to the total information as the degree of importance in discourse structure of text. Then I extracted top 50% and 10% important terms, and describe how they appear in text. The primary results are as follows: (1) new important terms tend to appear in former part of the text, and they are repeated constantly as a whole. (2) Low-frequent terms appeared in former part of the text are gradually repeated, while the number of low-frequent terms are increased when all terms are considered.
