ANNO: A General Annotation Tool for Bilingual Clinical Note Information Extraction

Area of Science:

• Natural Language Processing
• Medical Informatics
• Computational Linguistics

Background:

• The increasing demand for large-scale annotated clinical data is crucial for advancing machine learning models in healthcare.
• Developing efficient and accurate tools for bilingual complex clinical text annotation is essential for data preprocessing.

Observation:

• The ANNO tool was designed to facilitate human annotators in efficiently and accurately annotating clinical documents.
• Key features include a dictionary function for tagging word/phrase types, inter-annotator comparison for reconciliation, and automatic reflection of updated regular expression sets.

Findings:

• ANNO is a Docker-based web application, ensuring freedom from dependency issues and enabling easy user access.
• It supports sharing of annotation markups as dictionary-structured regular expression sets and allows cross-checking of annotated corpora among annotators.
• Standardized input (Microsoft Excel) and output (XML) formats streamline data handling.

Implications:

• ANNO is expected to significantly aid researchers in creating annotated clinical datasets for machine learning.
• The tool's features promote consistency and collaboration in the complex task of clinical text annotation.
• Facilitating the development of more robust and accurate AI models in the medical domain.