File size: 2,871 Bytes
fa224fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
---

dataset_info:
  features:
    - name: id
      dtype: string
    - name: poem_id
      dtype: int64
    - name: poet
      dtype: string
    - name: dynasty
      dtype: string
    - name: title
      dtype: string
    - name: text
      dtype: string
    - name: setiments
      dtype: string
  splits:
    - name: train
      num_bytes: 474283
      num_examples: 4000
    - name: test
      num_bytes: 127061
      num_examples: 1000
  download_size: 1865728
  dataset_size: 601344
configs:
  - config_name: default
    data_files:
      - split: train
        path: train-*.parquet
      - split: test
        path: test-*.parquet
---


# FSPC - Fine-grained Sentimental Poetry Corpus

FSPC is a manually-labelled Fine-grained Sentiment Poetry Corpus. Each poem and each line is annotated into 5 classes, namely negative, implicit negative, neutral, implicit positive and positive.

## Dataset Information

- **Version**: V1.0
- **Number of poems**: 5,000
- **Sentiment Classes**: 
  - 1: negative
  - 2: implicit negative
  - 3: neutral
  - 4: implicit positive
  - 5: positive

## Data Format

The dataset is stored in Parquet format with the following fields:

- `id`: Unique identifier (poem_id as string)

- `poem_id`: Poem sequence number
- `poet`: Poet name
- `dynasty`: Dynasty name (e.g., "唐", "宋")
- `title`: Poem title
- `text`: Full poem text with lines separated by newline characters (original "|" replaced with "
")
- `setiments`: JSON string containing sentiment labels in original format: `{"holistic": "1", "line1": "1", "line2": "1", "line3": "2", "line4": "2"}`

The dataset is split into train (80%) and test (20%) sets with stratified sampling based on holistic sentiment to ensure balanced distribution.

## Citation

If you use this corpus, please cite the following paper:

```

@inproceedings{chensentiment:19,

    author  = {Huimin Chen and Xiaoyuan Yi and Maosong Sun and Cheng Yang and Wenhao Li and Zhipeng Guo},

    title   = {Sentiment-Controllable Chinese Poetry Generation},

    year    = "2019",

    booktitle = {Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence},

    address = {Macao, China}  

}

```

## Original Dataset

The original dataset can be found in JSON format where each line is a poem with the following structure:

```json

{

    "poet": "韦庄", 

    "poem": "自有春愁正断魂|不堪芳草思王孙|落花寂寂黄昏雨|深院无人独倚门", 

    "dynasty": "唐", 

    "setiments": {"holistic": "1", "line1": "1", "line2": "1", "line3": "2", "line4": "2"}, 

    "title": "春愁"

}

```

Lines and sentiments in a poem are split by "|". The sentiments dict contains the holistic sentiment of whole poem, sentiment of the first line, ..., sentiment of the fourth line.