Back

ACAT: A Collaborative Platform for Efficient Aspect-Based Sentiment Dataset Annotation

Ana-Maria Luisa Mocanu, Ciprian-Octavian Truică, and Elena-Simona Apostol
National University of Science and Technology POLITEHNICA Bucharest, Romania

Abstract

Aspect-Based Sentiment Analysis requires high-quality, detailed datasets to train reliable machine learning models. However, existing annotation tools treat output as flat files, leaving researchers to manually consolidate multi-annotator data, reconstruct relational structures, and compute reliability metrics through custom scripts. This paper introduces ACAT, a web-based platform that bridges the gap betweenunstructured text and structured knowledge systems, natively supporting four ABSA workflows: (1) Aspect-Category Sentiment Analysis, (2) Clause-Level Segmentation, (3) Aspect-Term Sentiment Analysis with character-level position tracking, and (4) Aspect Sentiment Triplet Extraction with dual span offset preservation. Its core contribution is an automated ETL pipeline that aligns collaborative annotations and computes Inter-Annotator Agreement metrics directly at export, producing training-ready datasets without supplementary processing. Validated on 1,002 restaurant reviews, ACAT achieves a median annotation time of 31.58 seconds and a raw inter-annotator agreement of 0.78 to 0.86 in all tasks, demonstrating efficiency and consistent annotation quality across varying levels of annotator expertise.

Keywords: Aspect-Based Sentiment Analysis, Clause-Level Sentiment Analysis, Aspect-Term Sentiment Extraction, Aspect Sentiment Triplet Extraction, Data Curation, Knowledge Extraction

Platform Capabilities

Supported Workflows

Collaboration & Teams

Automated Metrics (IAA)

ETL Export Serialization

Interact with the controls below to see how ACAT parses, consolidates, and exports datasets natively into JSON, XML, and CSV. Note how Multi-Annotator exports automatically compute IAA metrics.