語彙頻度分析と因子分析による新聞投稿文の特徴把握

概要

論文の詳細を見る
Lately, the Japanese-language proficiency of university students has conspicuously been deteriorating. Given the situation, my students have been provided with some lessons in order to improve their command of Japanese. After the lessons, my students have to contribute their letters on daily life or current topics to the readers' column named "Koe (Voice)"of The Asahi Shimbun Osaka. The purpose of this paper is to provide my students with useful information that they can refer to in the creative writing practice. Presented in this paper are the results obtained by performing word frequency analysis and factor analysis of letters to the readers' column and "Watashi no Shiten (My Perspective)" reported in The Asahi Shimbun Osaka, and some discussions based on observation of the results. Word frequency analysis provides a list of the words that occur in a text and the number of times they occur; moreover, factor analysis is a process in which the values of observed data are expressed as functions of a number of possible causes to determine which are most important. In the analysis, I took full advantage of a digital database of articles of The Asahi Shimbun. Relevant articles were retrieved from the database by searching under a group of keywords such as "readers' column *Koe (Voice)* Osaka" and were analyzed by a morphological analysis system named "Wincha." Extracted from the morphemes obtained were only nouns and adjectives useful to identify characteristics of the articles. Following the morphological analysis, a decimal classification group was equivalently given to each morpheme extracted referring to "Word List by Semantic Principles" compiled by National Language Research Institute. The frequency of occurrence of each decimal classification group was cumulatively calculated and the distribution of the frequency was statistically analyzed. After the word frequency analysis, factor analysis of the articles was performed using SPSS for Windows that provides a powerful statistical analysis and data management system in a graphical environment. The decimal classification groups were set as the variables and maximum likelihood estimation was used as the statistical method. Selected were the check boxes of SPSS for Windows such as "KOM and Bartlett's test of sphericity," "Scree Plot," and "Promax" in the dialogue box of the rotation method. The word frequency analysis of letters to the readers' column indicated that the frequencies of the decimal classification groups of G4 and G7 were prominent in the distribution. G4 is a group of subject words related to family, friends, and so on; G7 is a group of nouns related to personal feelings and activities. The result explains that themes of the letters reported in the readers' column have genuine relevance to experiences or memories in daily life. On the other hand, the word frequency analysis of "Watashi no Shiten (My Perspective)" showed that the frequencies of the decimal classification groups of G8, G6, and G5 were conspicuous in the distribution. G8 is a group of nouns related to public feelings and activities; G6 is a group of nouns related to public places and organizations; G5 is a group of subject words related to public positions, occupational categories, and so on. The result conforms to a feeling that experts' opinions about various social problems tend to appear in "Watashi no Shiten (My Perspective)." The factor analysis of letters to the readers' column isolated two underlying factors. Factor 1 is collectively called relationships between readers and natural environments or products; Factor 2 is named relationships between readers and events in daily life. The factor analysis of "Watashi no Shiten (My Perspective)" separated two factors. Factor 1 is relevant to social problems; Factor 2 is closely connected with life or medical care. The results confirm the validity of the findings that are discovered as the results of the word frequency analysis. On the basis of the findings, I am going to provide our students with more modified materials for Japanese learning. Moreover, I will analyze sentence structures of letters reported in the readers' column to comprehensively understand their characteristics.
徳島文理大学の論文

語彙頻度分析と因子分析による新聞投稿文の特徴把握

スポンサーリンク

概要

著者

関連論文

スポンサーリンク