StructBert_finetuned_sentiment

StructBert_finetuned_sentiment Sentiment Classification

This model is a fine-tuned version of the Chinese StructBERT sentiment classification model for binary sentiment classification of management responses in Chinese earnings conference call Q&A.

Base Model

The model is initialized from the following Chinese StructBERT sentiment classification model:

https://modelscope.cn/models/iic/nlp_structbert_sentiment-classification_chinese-base

Task

Binary sentiment classification of management responses.

Labels

• 0: Negative
• 1: Positive

Model

StructBert_finetuned_sentiment is fine-tuned on the manually labeled dataset used in the associated research study.

For detailed information on the dataset, research design, training procedure, evaluation, and empirical results, please refer to the associated paper.

Source Code

The complete training and evaluation code is available in the associated GitHub repository https://github.com/SCUzqy24/Chinese-Earnings-Call-Sentiment/.

Acknowledgement

We thank the authors of the original StructBERT model and the ModelScope model contributors for making the pretrained model publicly available.