Identifying Sections in Scientific Abstracts using Conditional Random Fields

TLDR

Prior knowledge of abstract rhetorical structure aids text‑mining tasks such as extraction, retrieval, and summarization. This paper introduces a method to classify abstract sentences into objective, methods, results, and conclusions sections. The authors formulate the task as a sequential labeling problem and use Conditional Random Fields trained on automatically extracted Medline abstracts. The CRF approach achieved 95.5 % per‑sentence accuracy and 68.8 % per‑abstract accuracy, outperforming previous methods.

Abstract

OBJECTIVE: The prior knowledge about the rhetorical structure of scientific abstracts is useful for various text-mining tasks such as information extraction, information retrieval, and automatic summarization. This paper presents a novel approach to categorize sentences in scientific abstracts into four sections, objective, methods, results, and conclusions. METHOD: Formalizing the categorization task as a sequential labeling problem, we employ Conditional Random Fields (CRFs) to annotate section labels into abstract sentences. The training corpus is acquired automatically from Medline abstracts. RESULTS: The proposed method outperformed the previous approaches, achieving 95.5% per-sentence accuracy and 68.8% per-abstract accuracy. CONCLUSION: The experimental results showed that CRFs could model the rhetorical structure of abstracts more suitably.