README.md

metadata

license: apache-2.0
task_categories:
  - text-classification
task_ids:
  - multi-label-classification
language:
  - es
multilinguality:
  - monolingual
size_categories:
  - n<1K
pretty_name: DISCOPALPsychClassification
tags:
  - poetry
  - spanish
  - sonnet
  - psychology
  - multi-label-classification
  - disco
  - disco-pal
  - mteb
  - poetrymteb
  - embedding-evaluation
annotations_creators:
  - expert-generated
source_datasets:
  - DISCO PAL
  - DISCO
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: source_index
        dtype: int64
      - name: poem
        dtype: string
      - name: labels
        sequence: int64
      - name: label_names
        sequence: string
      - name: n_annotators
        dtype: int64
    splits:
      - name: train
        num_examples: 210
      - name: test
        num_examples: 55

DISCOPALPsychClassification

Multi-label psychological feature classification over Spanish sonnets for PoetryMTEB, derived from DISCO PAL (Barbado et al., 2020).

Experts marked whether each sonnet refers to literature-motivated psychological concepts (binary presence). Labels follow the DISCO PAL psychological inventory (rooted in personality-disorder / clinical-educational modelling literature cited by the authors), aggregated by median across three annotators.

Dataset Card

Item	Description
Source annotations	DISCO PAL (Processed Annotations/poems_corpus_all.csv)
Source poems	DISCO
Paper	Barbado et al., arXiv:2007.04626
Languages	Spanish (es)
Unit	Full sonnet text (poem)
Labels	Multi-label subset of 21 psychological concepts (binary)
Aggregation	Median of 3 experts; treat median ≥ 0.5 as positive
Filtering	Poems with no positive psychological label are removed (9 poems)
Size	train=210; test=55
Splits	Stratified by primary label (first taxonomy-ordered positive) ≈ 80% / 20%, seed=42
License	Apache License 2.0
Evaluation metrics	Multi-label classification on embeddings: macro/micro F1, Average Precision (AP)

Mean labels/poem ≈ 5.43.

Label taxonomy (21)

id	label_name (en)	es (source)	zh	train	test	total
0	Solitude	Soledad	孤独	50	13	63
1	Illusion	Ilusión	幻想/憧憬	59	14	73
2	Dreaming	Ensoñación	梦想/白日梦	36	10	46
3	Greatness	Grandiosidad	夸大/宏大感	81	24	105
4	Pride	Orgullo	骄傲	57	15	72
5	Irritability	Irritabilidad	易怒	29	7	36
6	Anxiety	Ansiedad	焦虑	61	15	76
7	Anger	Ira	愤怒	42	15	57
8	Instability	Inestabilidad	不稳定	55	9	64
9	Idealization	Idealización	理想化	84	23	107
10	Depression	Depresión	抑郁	30	9	39
11	Disappointment	Desilusión	失望	41	6	47
12	Dislike	Aversión	厌恶/反感	83	16	99
13	Insecurity	Inseguridad	不安全感	35	9	44
14	Impotence	Impotencia	无力感	49	13	62
15	Vulnerability	Vulnerabilidad	脆弱	99	30	129
16	Fear	Temor	畏惧	72	22	94
17	Obsession	Obsesión	强迫观念	28	4	32
18	Compulsion	Compulsión	强迫行为	47	9	56
19	Prejudice	Prejuicio	偏见	30	0	30
20	Dramatization	Dramatización	戏剧化	84	24	108

Codebook: label_taxonomy.json.

Upstream Spanish column names are preserved in the taxonomy (name_es) for traceability; classification uses English label_names.

Features

Field	Type	Description
id	string	Example id (discopal-psych-{source_index})
source_index	int64	Row index in upstream CSV
poem	string	Full Spanish sonnet text
labels	list[int64]	Psychological class indices
label_names	list[string]	Canonical English concept names
n_annotators	int64	Experts aggregated (3; median)

Construction method

1. Load poems_corpus_all.csv (median of a1/a2/a3).
2. Normalize column names; map 21 Spanish columns → English taxonomy.
3. Keep concepts with median ≥ 0.5; drop poems with an empty label set.
4. Stratified train/test split by primary (lowest-id positive) label, seed=42.

Intended use / limitations

• For PoetryMTEB multi-label probing of Spanish poem embeddings.
• Labels encode literary / conceptual presence, not clinical diagnoses of authors or readers.
• Companion affective dataset: DISCOPALAffectClassification.

Citation

@article{barbado2020disco,
  title={DISCO PAL: Diachronic Spanish Sonnet Corpus with Psychological and Affective Labels},
  author={Barbado, Alberto and Fresno, Víctor and Riesco, Ángeles Manjarrés and Ros, Salvador},
  journal={arXiv preprint arXiv:2007.04626},
  year={2020}
}

Underlying sonnets:

@misc{disco2017,
  author={Ruiz Fabo, Pablo and Bermúdez Sabel, Helena and Martínez Cantón, Clara and Calvo Tello, José},
  title={Diachronic Spanish Sonnet Corpus (DISCO)},
  year={2017},
  howpublished={UNED / Zenodo},
  doi={10.5281/zenodo.1069844},
  url={https://github.com/pruizf/disco}
}

Psychological inventory background (as cited by DISCO PAL):

@mastersthesis{garcia2016owl,
  author={García Franco, J. D. and Manjarrés Riesco, Ángeles},
  title={Modelado de los trastornos de la personalidad en OWL},
  school={UNED},
  year={2016}
}

License

Apache License 2.0 for DISCO PAL annotations and this redistribution. Poem texts originate from DISCO; respect upstream DISCO terms for uses beyond research evaluation.