| dc.contributor | 統計學系 | zh_Tw |
| dc.creator (作者) | 余清祥 | zh_TW |
| dc.date (日期) | 2015 | en_US |
| dc.date.accessioned | 26-Dec-2017 17:46:39 (UTC+8) | - |
| dc.date.available | 26-Dec-2017 17:46:39 (UTC+8) | - |
| dc.date.issued (上傳時間) | 26-Dec-2017 17:46:39 (UTC+8) | - |
| dc.identifier.uri (URI) | http://nccur.lib.nccu.edu.tw/handle/140.119/115435 | - |
| dc.description.abstract (摘要) | 多數國人中文字彙約一萬三千字，常用字則為五千餘字，大家知道及常用的字彙差異不大，但用字方式非常不同，亦即字彙的偏好或厭惡因人而異。如果把字彙視為生物物種，將每個人的寫作風格比擬為不同生態系統，用字習慣的改變則類似於生態變遷，而物種演化及生態結構的想法可用於比較寫作風格，以及探討風格變化的軌跡。本計畫以《新青年》十一卷及傳教士報刊為素材，結合生態變遷及統計方法比較各卷的異同，觀察語言轉換歷程，尋找可以建立客觀判讀文言文、白話文篇章的指標。本文以卷為分析單位，多角度量化《新青年》雜誌各卷的文本結構，包括文本用字、用句、文言和白虛字使用以及常用字詞共用等方面，通過多種圖表相結合的呈現方式，窺探《新青年》雜誌語言變化歷程以及轉變特點。這其中既包括了對文言文到白話文轉變機制的探索，也包括白話語言演化的探索。其次，根據各卷初探的結果，尋找可區隔文言文和白話文兩種語言形式的文本特徵變數，再以《新青年》第一卷和第七卷為訓練樣本，結合主成分和羅吉斯迴歸，對文、白兩種語言形式的文章進行分類訓練，再利用第四卷進行測試。結果證實，所提取的文本變數能夠有效實現對文、白兩種語言形式的文章的區分。此外，本文亦根據前述初探結果以及人文學者經驗，探索《新青年》雜誌後期語言形式的變化，即從五四運動時期的白話文至以「紅色中文」為特徵的白話文（二戰之後中國使用的白話文）的變化。以第七卷和第十一卷為樣本進行訓練，結果證實這兩卷語言形式存在明顯區別；並加入台灣《聯合報》和中國大陸《人民日報》進行分類預測，發現兩類報刊的語言偏向有明顯差異，值得後續深入研究。 | zh_TW |
| dc.description.abstract (摘要) | Many scholars have studied the language revolution in early Republic. Our team also contributed by bringing in digital humanity methods to find effective ways to tell the classical Chinese from modern Chinese by computer. It was when we analyzed the data of The New Youth magazine that we noticed something rarely discussed by modern scholars: the language changed significantly again in Volume 8 of the magazine. In this study, we adapt the idea of un-supervised learning from statistical learning theory to define and identify important variables. In particular, we will use the notion of Exploratory Data Analysis (or EDA, proposed by famous statistician J.W. Tukey in 1977) to evaluate potential variables which can differentiate the language styles of Volume 7 and Volumes 8~11 in New Youth Magazine. Our study shows that the styles of articles from UnitedDaily News are close to that of Volume 7 in The New Youth magazine, while those from People’s Daily have about equal probability for being classified to Volume 7 or to Volumes 8~11. This implies that the Chinese writing styles of Taiwan and mainland China are different and the writing style of mainland China is likely to be influenced by the Soviet language. | en_US |
| dc.format.extent | 651804 bytes | - |
| dc.format.mimetype | application/pdf | - |
| dc.relation (關聯) | 執行起迄：2015/08/01~2017/07/31 | zh_TW |
| dc.relation (關聯) | 104-2420-H-004-036-MY2 | zh_TW |
| dc.subject (關鍵詞) | 文字採礦;生物多樣性;《新青年》;探索性資料分析;相似指標 | zh_TW |
| dc.subject (關鍵詞) | Classical and Modern Chinese; New Youth Magazine;Unstructured Data; Exploratory Data Analysis; Logistic Regression | en_US |
| dc.title (題名) | 從文言文到白話文：由生態變遷探討語言變化 | _TW |
| dc.type (資料類型) | report | |
