Multi-label Text Classification of Economic Concepts from Economic News Articles using Natural Language Processing

Abstract

Multi-label classification is rapidly developing as an important aspect of modern predictive modeling. In this paper, we propose a multi-label text classification approach in order to extract the labels of economic concepts from economic news articles. We demonstrate a multi-label sentence-level event classification with a multi-label classifier algorithm. The classifier uses BERT Model and classification based on the association between labels via a threshold. The experiment on real-world multi-label data with many labels demonstrates an appealing performance and efficiency of multi-label classification.