中國研究資料庫與知識圖譜

The Knowledge Database/ Graph of "China-studies"

本研究建立臺灣學者的「中國研究論文資料庫」，其方法為蒐集「中國研究、大陸研究、中共」做為關鍵詞之「台灣岀版人社學術期刊」，共得1367篇。隨後，以其中「篇名、關鍵字、摘要」為基礎欄位（metadata）建立本資料庫。 上述期刊成果，富含中國研究的方方面面「知識」，能作為臺灣學者從事中國研究之「知識庫」。 本研究進一步使用隱含狄利克雷分布（Latent Dirichlet Allocation, LDA)演算法將各論文做分群，可發現臺灣學者的研究主題可分為「台商管理相關、經濟體制相關、國家政社理論、內政財政相關、傳播與法規範、政治意識型態、國際兩岸關係」等大項。 在本網站進一步可視化，將上述數據以社會網路分析（SNA）方法做出「年代-主題」、「作者-期刊」、「主題-關鍵字」等等中國研究知識圖譜，提供學者同儕作為後續研究基礎。 如有合作需求或意見回饋，請聯繫hlshao2@gmail.com

With rich accumulation of "China Studies" (CS) in Taiwan academia for many years and the concern for "Principle Development" recently, to “re-research CS" has gradually emerged. This article attempts to review Taiwan's "CS" using digital methods, including text mining (TM) and machine learning (ML) techniques. This study collected 1,367 papers published on the journal in Taiwan, from 1996 to 2019. Then we structured them, and established the "CS" database. This can successfully export the information of the author, the journal organization, and etc. for analysis. At the same time, this website attempts to cluster the topic form the "CS" database, which are: "Taiwanese business management, economic system related, state-politics-social theory, internal and financial related, communication and norms related, politics-ideology, international and cross-strait relations." This project also discusses these main CS topics, provides the keywords in each topic, and describes the trends over the years. The reaserch reached two goals: firstly, as a fundamental research, it, has tried to build the "CS" database. And secondly, as an applied research, it could locate and analyze into a specific the subgroup.

誌謝：本資料庫由科技部計畫「從知識系譜到知識地圖—大數據與機器學習下的中國研究」。
“From Knowledge Genealogy to Knowledge Map-China Studies in Big Data and Machine Learning”
107-2410-H-003 -058 -MY3, Ministry of Science and Technol-ogy, the MOST, Taiwan.

使用說明

我們的資料庫不是羅列關於期刊內容，而是進一步將其知識「體系化」。下列各欄位都可以自己輸入相關「作者/期刊/關鍵字」，可以看看有沒有熟悉的人事物。在右上角有各種「知識分頁」：
A. 在「知識圖譜群組」中，提供了「作者-期刊-關鍵字」的各種組合，讀者可以容易的找到「哪些學者都做什麼類型的研究」「他有哪些共通學術愛好的學者」。可以看看研究者們常投哪些期刊，以及他常常關心的議題。點點「顯示節點內容」，可以顯現更多資訊。
B. 在「作者統計/文章特性」中，讀者可以看到每個作者或文章是所有知識演變上的什麼時間、什麼位置，是使用數量式的呈現。可以很快知道每年有哪些論文。
C. 在「相似文本分析」中，你只要輸入某個關鍵字，我們會提供「相近關鍵字」與「相關文章」。我們使用了演算技巧，能夠提供近似概念，也可以接受不是完全一樣的字來「搜尋」。你也可以點選相似關鍵字（在右邊），可以擴展綿延相關知識。
D. 「主題分群分析」則是使用機器學習方式將其分群，可以知道「臺灣中國研究的主題」，知道歷年熱門的題目方向與關鍵字。

Readme

Our database does not list the content of the journal, but further "systematizes" its knowledge. You can enter the relevant "author/journal/keyword" in the "up-right" series webpages, and see if there are any familiar people and things in the research field.
A. In the "Groups of Knowledge Graph (知識圖譜群組)", various combinations of "author-journal-keywords" are provided. Readers can easily find "Which scholars do what kind of research" or "Which scholars do they have in common academic hobbies" . You can look at which journals researchers often submit to, and the topics that he often cares about. Click "Show node content" to display more information.
B. In "Author Statistics/Article Features (作者統計/文章特性)", you can see when and where each author or article is in the evolution of all knowledge, which is presented in quantitative. You can quickly know how many papers are available each year.
C. In "Similar Text Analysis (相似文本分析)", you only need to enter a certain keyword, and we will provide "Similar Keywords" and "Related Articles". We have used machine-learning techniques to provide approximate concepts, and we can also "search" by accepting words that are not exactly the same words. You can also click on similar keywords (on the right) to expand and extend relevant knowledge.
D. In "Subject cluster analysis (主題分群分析)",we use also machine learning methods to group them, so that you can know the "topics of Taiwanese Chinese research" and the direction of current popular topics and its keywords.