Determining Sentiment Intensity of English and Arabic Phrases
Determine the sentiment intensity of English and Arabic phrases using Best-Worst Scaling: annotators see four phrases at a time and select the most positive and most negative, which are aggregated into real-valued intensity scores in [0,1]. Based on SemEval-2016 Task 7.
Configuration Fileconfig.yaml
This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.
# Determining Sentiment Intensity of English and Arabic Phrases
# Based on Kiritchenko et al., SemEval 2016
# Paper: https://aclanthology.org/S16-1004/
# Dataset: http://saifmohammad.com/WebPages/BestWorst.html
#
# This task uses Best-Worst Scaling (BWS, a.k.a. MaxDiff): annotators are shown
# a tuple of four phrases and select the ONE that is most positive (best) and the
# ONE that is most negative (worst). Aggregating many such judgments yields a
# real-valued sentiment intensity score in [0, 1] for each phrase (0 = most
# negative, 1 = most positive).
annotation_task_name: "Sentiment Intensity (Best-Worst Scaling)"
task_dir: "."
data_files:
- sample-data.json
item_properties:
id_key: "id"
text_key: "prompt"
output_annotation_dir: "annotation_output/"
output_annotation_format: "json"
port: 8000
server_name: localhost
annotation_schemes:
- annotation_type: radio
name: bws_best
description: "Which phrase is the MOST POSITIVE?"
horizontal: true
labels:
- "Phrase 1"
- "Phrase 2"
- "Phrase 3"
- "Phrase 4"
sequential_key_binding: true
- annotation_type: radio
name: bws_worst
description: "Which phrase is the MOST NEGATIVE?"
horizontal: true
labels:
- "Phrase 1"
- "Phrase 2"
- "Phrase 3"
- "Phrase 4"
sequential_key_binding: true
annotation_instructions: |
You will be shown a set of four phrases at a time. For this set, make two
judgments:
- Best: pick the ONE phrase that is the MOST POSITIVE in sentiment.
- Worst: pick the ONE phrase that is the MOST NEGATIVE in sentiment.
The "best" and "worst" phrases must be different. These comparative choices are
later aggregated (Best-Worst Scaling) into a real-valued sentiment intensity
score in [0, 1] for each phrase.
html_layout: |
<div style="padding: 15px; max-width: 800px; margin: auto;">
<div style="background: #f0f9ff; border: 1px solid #bae6fd; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
<strong style="color: #0369a1;">Compare these four phrases:</strong>
<p style="font-size: 18px; line-height: 1.7; margin: 8px 0 0 0;">{{prompt}}</p>
</div>
</div>
allow_all_users: true
instances_per_annotator: 50
annotation_per_instance: 2
allow_skip: true
skip_reason_required: false
Sample Datasample-data.json
[
{
"id": "bws_001",
"prompt": "Phrase 1: happy accident | Phrase 2: total failure | Phrase 3: pretty good | Phrase 4: mildly annoying",
"options": [
"happy accident",
"total failure",
"pretty good",
"mildly annoying"
]
},
{
"id": "bws_002",
"prompt": "Phrase 1: absolutely wonderful | Phrase 2: not bad | Phrase 3: deeply disappointing | Phrase 4: painfully boring",
"options": [
"absolutely wonderful",
"not bad",
"deeply disappointing",
"painfully boring"
]
}
]
// ... and 6 more itemsGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git cd potato-showcase/semeval/2016/task07-sentiment-intensity potato start config.yaml
Dataset & paper
Kiritchenko et al., SemEval 2016
Citation (BibTeX)
@inproceedings{kiritchenko-etal-2016-semeval,
title = "{S}em{E}val-2016 Task 7: Determining Sentiment Intensity of {E}nglish and {A}rabic Phrases",
author = "Kiritchenko, Svetlana and Mohammad, Saif M. and Salameh, Mohammad",
booktitle = "Proceedings of the 10th International Workshop on Semantic Evaluation ({S}em{E}val-2016)",
year = "2016",
address = "San Diego, California",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/S16-1004",
doi = "10.18653/v1/S16-1004",
pages = "42--51"
}Details
Annotation Types
Domain
Use Cases
Tags
Found an issue or want to improve this design?
Open an IssueRelated Designs
Aspect-Based Sentiment Analysis
Identification of aspect terms in review text with sentiment polarity classification for each aspect. Based on SemEval-2016 Task 5 (ABSA).
Fine-Grained Sentiment Analysis on Financial Microblogs and News
Graded sentiment analysis of financial text with topic classification, rating market sentiment from very bearish to very bullish on a 7-point scale. Based on SemEval-2017 Task 5.
AdMIRe - Advancing Multimodal Idiomaticity Representation
Multimodal idiomaticity task (SemEval-2025 Task 1, Subtask A): given a potentially idiomatic nominal compound in a context sentence, annotators rank five candidate images by how well each represents the sense in which the compound is used.