README.md

metadata

license: cc0-1.0
task_categories:
  - text-classification
task_ids:
  - multi-class-classification
language:
  - zh
multilinguality:
  - monolingual
size_categories:
  - 1K<n<10K
pretty_name: Classical Chinese Poetry Theme Classification
tags:
  - poetry
  - chinese
  - classical-chinese
  - theme-classification
  - multi-class-classification
  - mteb
  - poetrymteb
  - embedding-evaluation
annotations_creators:
  - expert-generated
source_datasets:
  - doi:10.7910/DVN/6QJ7RK
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
    default: true
dataset_info:
  - config_name: default
    features:
      - name: id
        dtype: string
      - name: poem
        dtype: string
      - name: label
        dtype: int64
      - name: label_name
        dtype: string
      - name: label_name_en
        dtype: string
      - name: label_name_zh
        dtype: string
    splits:
      - name: train
        num_examples: 2624
      - name: test
        num_examples: 292

Classical Chinese Poetry Theme Classification

Multi-class theme classification of classical Chinese poetry for PoetryMTEB embedding evaluation.

Upstream: A thematic classification dataset for Classical Chinese poetry (Hou, Harvard Dataverse, DOI), with Hou & Zhang, JOCH 2024 (paper). Nearly 3,000 annotated poems across seven themes (Chunqiu → Qing).

Dataset Card

Item	Description
Dataset version (PoetryMTEB)	1.0.1
Source	Harvard Dataverse DVN/6QJ7RK (train.xlsx / test.xlsx)
Paper	Hou J, Zhang S. Exploring Thematic Diversity in Classical Chinese Poetry… JOCH 17(4), 2024 (DOI)
Languages	Classical Chinese / Chinese (zh)
Unit	Full poem text (poem)
Labels	Single-label 7 themes (label 0–6, matching Dataverse README)
Size	train=2624; test=292; total=2916
Splits	train / test only (no validation)
License	CC0 1.0
Evaluation metrics	Accuracy, macro/micro F1

Label taxonomy (7)

id	label_name (zh)	name_en	train	test	total
0	爱情婚姻	Love and Marriage	269	43	312
1	交友送别	Friendship and Farewell	621	74	695
2	羁旅思乡	Journey and Homesickness	318	35	353
3	边塞战争	Frontier and War	207	23	230
4	山水田园	Landscape and Countryside	337	31	368
5	咏史怀古	History and Nostalgia	354	33	387
6	托物言志	Expressing emotions through objects	518	53	571

id	Description (EN)
0	Joys, sorrows, and ups and downs of love, marriage, and family life.
1	Praise of friendship, sorrow of parting, anticipation of reunion.
2	Longing for hometown and family; loneliness of travel and exile.
3	Border warfare; loyalty, courage, and sacrifice of soldiers.
4	Love of natural scenery; pastoral and landscape imagery.
5	Historical events, figures, and cultural memory; nostalgia.
6	Thoughts and feelings conveyed via portrayal/symbolism of objects.

Codebook: label_taxonomy.json.

Features

Field	Type	Description
id	string	Example id (ccptc-{split}-{row})
poem	string	Classification input
label	int64	Theme id 0–6
label_name	string	Chinese theme name
label_name_en	string	English theme name
label_name_zh	string	Chinese theme name

Construction method

1. Download upstream train.xlsx / test.xlsx from Dataverse.
2. Keep official 0–6 labels; attach ZH/EN names from the Dataverse README.
3. Write PoetryMTEB parquet + Dataset Card + supporting materials.

Supporting materials

Path	Content
VERSION / LICENSE / CITATION.* / metadata.json	Packaging metadata
label_taxonomy.json	Theme codebook (ZH/EN)
splits/*_ids.txt	Split id lists
checksums.sha256	File checksums
scripts/evaluate_cc_theme_clf.py	Baseline TF-IDF + LR probe

How to load

from datasets import load_dataset

ds = load_dataset("PoetryMTEB/ClassicalChinesePoetryThemeClassification")
print(ds["train"][0]["poem"][:80])
print(ds["train"][0]["label_name"], ds["train"][0]["label_name_en"], ds["train"][0]["label"])

Citation

@article{10.1145/3685679,
  author  = {Hou, Jingrui and Zhang, Shitou},
  title   = {Exploring Thematic Diversity in Classical Chinese Poetry: A Novel Dataset and a BERT-enhanced Ensemble Learning Approach},
  year    = {2024},
  journal = {Journal on Computing and Cultural Heritage},
  volume  = {17},
  number  = {4},
  publisher = {ACM},
  doi     = {10.1145/3685679},
  url     = {https://doi.org/10.1145/3685679}
}

@data{hou2024classical_theme,
  author = {Hou, Jingrui},
  title  = {A thematic classification dataset for Classical Chinese poetry},
  year   = {2024},
  publisher = {Harvard Dataverse},
  doi    = {10.7910/DVN/6QJ7RK},
  url    = {https://doi.org/10.7910/DVN/6QJ7RK}
}

This Hub packaging: PoetryMTEB/ClassicalChinesePoetryThemeClassification (version 1.0.1)

License

CC0 1.0 (Dataverse).

---

古典汉语诗歌主题分类（中文说明）

7 类：爱情婚姻、交友送别、羁旅思乡、边塞战争、山水田园、咏史怀古、托物言志。
train=2624 / test=292（无 validation）；标签 0–6。