theoracle's picture
Update README.md
ac701d3 verified
|
raw
history blame
1.46 kB
metadata
license: apache-2.0
task_categories:
  - sentiment-analysis
tags:
  - sentiment
  - italian
  - news headlines
size_categories:
  - n<1K
Dataset Description:
  General Description: >-
    This dataset consists of Italian news headlines with annotated sentiments.
    Each headline is enclosed in square brackets followed by the sentiment label
    'positive', 'neutral', or 'negative'.
  Purpose: >-
    The dataset is designed for training and evaluating sentiment analysis
    models on Italian-language text, particularly news headlines.
Dataset Structure:
  Size and Scope: >-
    The dataset contains a small number of annotated headlines, suitable for
    initial model training or testing in sentiment analysis tasks.
  Data Fields: Each record includes a 'headline' text field and a 'sentiment' label.
  Example:
    - headline: '[ mi fa sbagliare tutte le paroleeeee.]'
      sentiment: negative
    - headline: '[ perfetto hai visto poi alla fine anche oggi e passato..]'
      sentiment: neutral
    - headline: >-
        [Rutelli: appoggio al governo #monti, sta lavorando bene #ballarò
        #osservatoriotivvù]
      sentiment: positive
Use Cases:
  Sentiment Analysis Model Training: >-
    Researchers and practitioners can use this dataset to develop and train
    sentiment analysis models for the Italian language.
  Academic Research: >-
    The dataset can serve as a basis for studies in computational linguistics
    focusing on sentiment analysis in Italian news media.