Last updated: Sep 26, 2026最後更新：2026年9月26日

CHI 3242 Course SyllabusCHI 3242 課程大綱

Text Mining for Chinese Humanities
中國人文研究中的文本挖掘
Term 1, 2026–20272026–2027 學年第一學期

Seminar:研討課： Thu, 3:30–6:30 pm (LCH 213); 17 Sep cancelled; makeup 21 or 25 Sep, 09:00–11:00, LCH 202 (choose one)星期四，下午 3:30–6:30（LCH 213）；9月17日停課；補課於 9月21日或 9月25日，09:00–11:00，LCH202（二選一）

Instructor:教師： Maciej Kurzynski 馬傑 (maciej.kurzynski[at]ln.edu.hk)
Office hours:辦公室時間： Tue 3:00–5:00 pm and Fri 10:00 am–12:00 pm, HSH G06, or by appointment星期二下午 3:00–5:00 及星期五上午 10:00–12:00，HSH G06，或另行預約
Language of Instruction:授課語言： Putonghua（普通話）普通話

Illustration for Text Mining for Chinese Humanities

🛠️ New to agentic coding? Start here: How to Set Up an AI Coding Agent. Please complete this before the first seminar.

🛠️ 初次接觸智能體編程？請先看：如何安裝人工智能編程助手。請於第一次研討課之前完成。

Course Description課程簡介

This course offers a beginner-friendly introduction to Digital Humanities, surveying tools for exploring and analyzing Chinese texts, including novels, blogs, news articles, poetry, and more. Students will use accessible software and learn basic techniques to study patterns in Chinese literature and social media, assess trends in digital archives, and visualize data. Prior programming experience is not required. From early in the course we will use agentic coding: AI coding assistants that help write, debug, and evaluate research code. Weekly seminars cover fundamental concepts and methods and provide supervised hands-on practice. For the final project, students may use corpora shared in class or build their own datasets, apply digital methods to a topic of their choice, and present their findings through a complete working GitHub repository and an accompanying report.

本課程以淺白方式介紹數字人文（Digital Humanities），廣泛概覽可用於探索與分析中文文本（小說、網誌、新聞、詩歌等）的數字工具。學生將運用易於上手的軟件，學習基本技術，以研究中國文學與社交媒體中的模式、評估數字檔案中的趨勢，並進行數據可視化。無須具備編程經驗。課程初期即引入智能體編程（agentic coding）：以人工智能編程助手協助撰寫、除錯並評估研究代碼。每週研討課講授數字人文的基本概念與方法，並在教師指導下進行實作練習。期末專題中，學生可使用課堂提供的語料，或自建數據集，將數字方法應用於自選題目，並以完整可運作的 GitHub 倉庫及配套報告呈現研究成果。

Aims課程目標

This course aims to introduce students to digital humanities in both practical and theoretical terms. Students will develop digital literacy, including the ability to employ digital tools and AI coding agents in creative ways, to evaluate information on the web critically, and to manage that information by searching, editing, uploading, and visualizing textual data. Familiarity with these tools equips students with transferable skills and with the means to champion more inclusive narratives and a broader cultural understanding.

本課程旨在從實踐與理論兩方面，引介學生進入數字人文領域。學生將培養數字素養，包括創造性地運用數字工具與人工智能編程助手、批判地評估網絡資訊，以及透過搜尋、編輯、上傳與可視化文本數據來管理資訊。熟悉這些工具，將使學生掌握可遷移的技能，並有能力推動更具包容性的敘事與更廣的文化理解。

Learning Outcomes學習成果

Upon completion of this course, students should be able to:

  • Compare different tools used for text mining and data visualization
  • Use simple Python scripts, packages, and AI coding agents to analyze literary phenomena
  • Detect biases in digital archives using basic probability
  • Identify research questions that digital humanities methods can address
  • Design and complete a digital humanities project

完成本課程後，學生應能：

  • 比較用於文本挖掘與數據可視化的不同工具
  • 運用簡易的 Python 程式、套件與人工智能編程助手分析文學現象
  • 以基本機率方法偵測數字檔案中的偏差
  • 辨識可用數字人文方法探討的研究問題
  • 設計並完成一項數字人文專題

Language of Instruction授課語言

The course is conducted in Putonghua. Primary materials (novels, poetry, and other Chinese texts) are in Chinese. Technical readings and Python documentation are often in English, and you are expected to read them in English unless a Chinese translation is available. The instructor might provide translated versions if needed. There are many AI tools that can help you translate technical texts free of charge (e.g., Perplexity, DeepSeek, Poe, Gemini). In-class discussions will be held in Putonghua, and students should complete and submit all assessed work in Chinese.

本課程以普通話進行。一手材料（小說、詩歌及其他中文文本）均為中文。技術文本與 Python 文獻多為英文，除非已有中譯，否則須以英文閱讀。如有需要，教師或會提供譯本。亦有不少免費人工智能工具可協助翻譯技術文本（例如 Perplexity、DeepSeek、Poe、Gemini）。課堂討論以普通話進行，所有評核作業均須以中文完成並提交。

Preliminary Setup課前準備

Please complete the following before the first seminar:

  1. Create a GitHub account.
  2. Collect your API key at keyreg.qhchina.org (class code: chi3242). Do not put the key in a repository or share it with others.
  3. A classroom with computers will be available. You may also bring your own laptop. If you use your own computer, install Python, Visual Studio Code (VS Code), and then OpenCode before the first class (or use GitHub Codespaces instead, if you cannot install locally). 👉 Step-by-step guide: How to Set Up an AI Coding Agent.

We will learn agentic coding early in the course and use AI coding assistants in VS Code and OpenCode to write and debug research code.

請於第一次研討課之前完成下列準備：

  1. 註冊 GitHub 帳號。
  2. 請到 keyreg.qhchina.org 領取你的 API 金鑰（課程代碼：chi3242）。不要把金鑰寫進倉庫，也不要與他人分享。
  3. 課堂備有電腦室。亦可自備手提電腦。若使用個人電腦，請於開課前安裝 Python、Visual Studio Code（VS Code），然後安裝 OpenCode（若無法本機安裝，可改用 GitHub Codespaces）。👉 詳細步驟：如何安裝人工智能編程助手。

課程初期即講授智能體編程，並在 VS Code 與 OpenCode 中使用人工智能編程助手撰寫與除錯研究代碼。

Assessment評核

Students’ progress towards the learning outcomes will be measured by means of:

ComponentWeight
In-class discussion and participation20%
Assignments (2)30%
Project proposal10%
Final project40%

Grading scale: A 85%–100%; A− 80%–84%; B+ 75%–79%; B 70%–74%; B− 65%–69%; C+ 60%–64%; C 55%–59%; C− 50%–54%; D+ 45%–49%; D 40%–44%; F 0%–39%

學生的學習成果將以下列方式評核：

項目比重
課堂討論與參與20%
作業（2份）30%
專題計劃書10%
期末專題40%

評分等級： A 85%–100%；A− 80%–84%；B+ 75%–79%；B 70%–74%；B− 65%–69%；C+ 60%–64%；C 55%–59%；C− 50%–54%；D+ 45%–49%；D 40%–44%；F 0%–39%

Attendance Policy出席規定

Students are expected to attend all seminars. One unexplained absence is permitted, no questions asked. Unexplained absences will not impact the final grade by more than 20% (the maximum you can get for in-class discussion and participation), unless a student misses more than four (4) meetings, in which case there will be insufficient evidence of participation to award a passing grade. Every unexplained absence (beyond the one everyone gets for free) will remove 10% from the participation component (which equals 2% of the final grade). For example, if you are absent two times without explanation (in addition to the one no-questions-asked absence granted to everyone), you will receive 100% − 20% = 80% of the total possible contribution from participation, which amounts to 80% × 20% = 16% of the final grade. Remember to come to class on time and prepared.

學生須出席所有研討課。一次無故缺席不計，無須解釋。無故缺席對學期成績的影響不會超過 20%（即課堂討論與參與的上限），惟缺席超過四次者，將因缺乏足夠評核依據而不能獲得及格成績。在人人皆有的一次免費缺席之外，每次無故缺席將從參與分扣除 10%（即學期總分的 2%）。例如，若除人人皆有的一次外再無故缺席兩次，參與分可得 100% − 20% = 80%，即學期總分的 80% × 20% = 16%。請準時出席，並預先準備。

AI Policy人工智能使用規定

This course teaches agentic coding, and students are expected to use AI coding assistants. Every such use must be explicitly documented. Any assignment completed with the assistance of an AI tool must include a statement identifying the tool and explaining how it was employed. Students must understand and be able to explain the code and analysis they submit. AI-generated prose may not be reproduced in submitted work, whether verbatim or with only superficial modifications. Doing so constitutes academic misconduct and will affect the assessment of the work.

本課程講授智能體編程，學生須使用人工智能編程助手。任何此類使用均須明確記錄。凡使用人工智能工具協助完成的作業，均須附上聲明，註明所使用的工具，並說明其具體用途。學生必須理解並能解釋所提交的程式碼與分析。不得在所提交的作業中原文照錄人工智能生成的分析文字，亦不得僅作表面修改後使用。違者將被視為學術不當行為，並會影響該作業的評核結果。

Course Design課程設計

1. Seminars. Each week, a three-hour seminar combines conceptual discussion with supervised hands-on practice. Students work with Chinese texts and learn methods ranging from word frequencies and collocations to topic models, word vectors, and neural language models. Prior programming experience is not required. We introduce agentic coding in the first weeks and then use AI coding agents throughout.

2. Assignments and project proposal. Two graded assignments apply the tools covered in seminars. Assignment 0 is not graded. The final assignment item is a final project proposal. All work must be submitted via the Assignment module on Moodle. Instructions are linked in the schedule table.

3. Final project. Students may use corpora shared in class or build their own datasets, apply digital methods to a topic of their choice, and present their findings through a complete working GitHub repository (sources, code, and README files) and an accompanying report in Chinese. Submit the repository URL and the report through Moodle before the deadline. The report should include the following parts:

  • Abstract: a summary of goals, experiments, and findings (max. 150 characters)
  • Dataset: corpus description, sources, and preprocessing steps
  • Methods: technique(s) used and the rationale; a brief explanation of how the method works
  • Analysis: results, visualizations, interpretations, and theoretical discussion
  • Limitations: challenges, biases, and possible improvements

1. 研討課：每週三小時研討課結合概念講授與教師指導下的實作。學生處理中文文本，學習從詞頻、搭配、主題模型、詞向量到神經語言模型等方法。無須具備編程經驗。最初數週引入智能體編程，其後貫穿全學期使用人工智能編程助手。

2. 作業與專題計劃書：共兩份計分作業；作業0不計分。最後一項作業為期末專題計劃書。一律經 Moodle 作業區繳交。作業說明連結見下方課程進度表。

3. 期末專題：學生可使用課堂提供的語料，或自建數據集，將數字方法應用於自選題目，並以完整可運作的 GitHub 倉庫（含原始資料、程式碼與 README）及中文配套報告呈現研究成果。於截止日期前經 Moodle 提交倉庫網址及報告。報告須包括以下部分：

  • 摘要：目標、實驗與發現的概述（最多 150 字）
  • 數據集：語料說明、來源與預處理步驟
  • 方法：所用技術及其理據；簡要說明方法如何運作
  • 分析：結果、可視化、詮釋與理論討論
  • 局限：挑戰、偏差與可能的改進

Class Schedule 課程進度

Week週次 Date日期 Topic課題 Readings閱讀 Notes
1 3 Sep9月3日 Introduction導論
  • Digital Humanities: key concepts; agentic coding; Git, GitHub, OpenCode, and VS Code數字人文：核心概念；智能體編程；Git、GitHub、OpenCode 與 VS Code
 Notes
2 10 Sep9月10日 Language, Text, Encoding語言、文本與編碼 Notes
17 Sep9月17日 No class (cancelled). Makeup next week — choose one session: Mon 21 Sep or Fri 25 Sep, 09:00–11:00, LCH 202. The makeup covers Evert / collocations. 停課。下週補課，請二選一：9月21日（一）或 9月25日（五），09:00–11:00，LCH202。補課內容為 Evert 搭配詞。
3 21 Sep (Mon)
25 Sep (Fri)9月21日（一）
9月25日（五）		 Collocations (makeup)搭配詞（補課）
  • Choose one: 21 Sep (Mon) or 25 Sep (Fri), 09:00–11:00, LCH 202二選一：9月21日（一）或 9月25日（五），09:00–11:00，LCH202
  • Collocation as a method; the contingency table; statistical significance, Fisher's exact test, p value搭配作為方法；列聯表；統計顯著性、Fisher 精確檢定、p 值
  • Stefan Evert, “Corpora and Collocations”
  • (optional)（選讀） Paul Baker, Using Corpora in Discourse Analysis (2006), Chapter 5, “Collocations”
  • (optional)（選讀） J. R. Firth, “A Synopsis of Linguistic Theory” (1957), the “company it keeps” passage
 Notes
Slides
23 Sep9月23日 Assignment 0 (not graded) due, 9:00 am作業0（不計分）截止，上午9:00
4 24 Sep9月24日 What Is a Character? (Woloch)什麼是「角色」？（Woloch）
  • Character, word, space角色、詞語、空間
  • Alex Woloch, The One vs. the Many, pp. 12–42
  • Franco Moretti, “Network Theory, Plot Analysis” (Stanford Literary Lab Pamphlet 2)
  • (optional)（選讀） Andrew Piper, Enumerations, chapter on characterization
 Notes
5 1 Oct10月1日 No class (National Day)停課（國慶日）
6 8 Oct10月8日 From Words to Corpora從詞語到語料庫 Notes
9 Oct10月9日 Assignment 1 due, 9:00 am作業1截止，上午9:00
7 15 Oct10月15日 Topic Modeling主題模型
  • LDA; what is a “topic”?LDA；什麼是「主題」？
  • 蘇祺：〈主題模型〉，第十章，李飛躍、桑海（主編）《數字人文引論》。北京：中華書局，2026年
  • (optional)（選讀） Ted Underwood, “Topic Modeling Made Just Simple Enough” (2012)
  • (optional)（選讀） David M. Blei, “Probabilistic Topic Models” (2012)
 Notes
Slides
8 22 Oct10月22日 Stylometry文體計量學 Slides
Z-scores
9 29 Oct10月29日 Classification I分類（一）
  • Logistic regression for text classification (sentiment analysis)用於文本分類的邏輯迴歸（情感分析）
  • Binary and multi-class classification (softmax); loss function (cross-entropy)二元與多類分類（softmax）；損失函數（交叉熵）
  • TF-IDF
  • Daniel Jurafsky and James H. Martin, Speech and Language Processing (3rd ed. draft), Chapter 4, “Naive Bayes and Sentiment Classification” (sections 4.1–4.4)
  • (optional)（選讀） Hoyt Long and Richard Jean So, “Literary Pattern Recognition: Modernism between Close Reading and Machine Learning,” Critical Inquiry 42, no. 2 (2016)
4 Nov11月4日 Assignment 2 due, 9:00 am作業2截止，上午9:00
10 5 Nov11月5日 Encoders I編碼器（一）
11 12 Nov11月12日 Encoders II + Classification II編碼器（二）＋分類（二）
  • Neural networks神經網絡
  • Training BERT-based classifiers (sentiment analysis); uploading a model to Hugging Face訓練基於 BERT 的分類器（情感分析）；上載模型至 Hugging Face
  • Kurt Vonnegut, “The Shapes of Stories” (lecture)（講座）
  • Jay Alammar, “The Illustrated BERT”
  • (optional)（選讀） Katherine Elkins, The Shapes of Stories: Sentiment Analysis for Narrative
  • (optional)（選讀） Andrew Reagan et al., “The Emotional Arcs of Stories Are Dominated by Six Basic Shapes”, EPJ Data Science (2016)
  • (optional)（選讀） 3Blue1Brown, neural-network series (video)（影片）
18 Nov11月18日 Final Project Proposal due, 9:00 am期末專題計劃書截止，上午9:00
12 19 Nov11月19日 Decoders and LLMs解碼器與大型語言模型
  • From BERT to GPT; generative models; what is “context”?; next-token probability從 BERT 到 GPT；生成模型；什麼是「上下文」？；下一詞元機率
  • LLMs for data extraction; Python API (OpenRouter)以大型語言模型抽取資料；Python API（OpenRouter）
  • Jay Alammar, “The Illustrated GPT-2”
  • (optional)（選讀） Andrej Karpathy, “Intro to Large Language Models” (video)（影片）
  • (optional)（選讀） Emily M. Bender et al., “On the Dangers of Stochastic Parrots” (FAccT 2021)
13 26 Nov11月26日 Review and Loose Ends課程總結與收尾
  • Integrative review of methods, tying up loose ends, and final-project preparation方法整合回顧、補足未竟內容，以及期末專題準備
  • Miriam Posner, “How Did They Make That?”
  • (optional)（選讀） Nan Z. Da, “The Computational Case against Computational Literary Studies,” Critical Inquiry 45, no. 3 (2019)
7 Dec12月7日 Final project due, 9:00 am期末專題截止，上午9:00

Important Notes 注意事項

  1. Students are expected to spend a total of 9 hours (i.e. 3 hours of class contact and 6 hours of personal study) per week to achieve the course learning outcomes.
  2. Students shall be aware of the University regulations about dishonest practice in course work, tests and examinations, and the possible consequences as stipulated in the Regulations Governing University Examinations. In particular, plagiarism, being a kind of dishonest practice, is “the presentation of another person’s work without proper acknowledgement of the source, including exact phrases, or summarised ideas, or even footnotes/citations, whether protected by copyright or not, as the student’s own work.” Students are required to strictly follow university regulations governing academic integrity and honesty.
  3. To enhance students’ understanding of plagiarism, a mini-course “Online Tutorial on Plagiarism Awareness” is available on https://pla.ln.edu.hk/.
  1. 學生每週須投入共 9 小時（即 3 小時課堂接觸、6 小時自修），以達致課程學習成果。
  2. 學生須注意大學有關習作及考試不誠實行為的規例，以及《大學考試規例》所訂的可能後果。抄襲屬不誠實行為，指「未經妥為註明來源，將他人作品（包括原句、撮述的觀點，以至腳註／引文，無論是否受版權保護）呈交作為自己的作品」。學生須嚴格遵守大學有關學術誠信的規例。
  3. 為加深對抄襲的認識，可修讀小型課程「Online Tutorial on Plagiarism Awareness」：https://pla.ln.edu.hk/。