A Deep Learning-Based Approach to Constructing a Domain Sentiment Lexicon: a Case Study in Financial Distress Prediction

作者：

Highlights：

• Proposing a novel lexicon generating approach for sentiment analysis that integrates word embedding models with deep learning-based classifiers.

• Constructing financial domain sentiment lexicon in Chinese context is significant for analyzing related financial issues in China.

• Chinese financial domain sentiment lexicon (CFDSL) generated by this study contains four aspects of sentiment words, namely capital markets, stock markets, companies’ internal business conditions, and politics.

• Experiments prove that sentiment features extracted through CFDSL can independently achieve relatively satisfactory predictive performance in terms of financial distress prediction.

摘要

•Proposing a novel lexicon generating approach for sentiment analysis that integrates word embedding models with deep learning-based classifiers.•Constructing financial domain sentiment lexicon in Chinese context is significant for analyzing related financial issues in China.•Chinese financial domain sentiment lexicon (CFDSL) generated by this study contains four aspects of sentiment words, namely capital markets, stock markets, companies’ internal business conditions, and politics.•Experiments prove that sentiment features extracted through CFDSL can independently achieve relatively satisfactory predictive performance in terms of financial distress prediction.

论文关键词：Domain sentiment lexicon,Financial text mining,Deep learning,Financial distress prediction,Word vector

论文评审过程：Received 7 December 2020, Revised 12 May 2021, Accepted 27 June 2021, Available online 13 July 2021, Version of Record 13 July 2021.

论文官网地址：https://doi.org/10.1016/j.ipm.2021.102673