metadata

license: mit
task_categories:
  - text-classification
language:
  - en
tags:
  - emotion
  - complexity
  - readability
  - sentiment
pretty_name: CAMEO
size_categories:
  - 10K<n<100K

Dataset Card for CAMEO

Dataset to accompany the EMNLP'23 paper titled: "Misery Loves Complexity: Exploring Linguistic Complexity in the Context of Emotion Detection".

Dataset Details

50,000 subset from the GoEmotions Dataset automatically annotated with the following linguistic complexity measures:

idt: Incomplete Dependency Theory
dlt: Dependency Locality Theory
nnd: Nested-Nouns Distance
le: Left-embededness
percentage_polysyllable_words: % of polysyllable words
avg_conn_doc: Average connectives per sentence
number_of_uniq_entities: Number of unique named entities
average_word_len: Average word length
dale_word_frequency_score: DALE Word Frequency Score
avgtfidf: Average TF-IDF of all words based on the background corpus
avgll: Average Log-likelihood of all words based on the background corpus
type_token_ratio_perc: % Type-token ratio

Please refer to the paper for further details on the metrics or other information.

For details on how the data was collected or annotated for emotions. Please refer to the original GoEmotions dataset.