ESG analysis can help investors determine a business' long-term sustainability and identify associated risks. FinBERT-esg-9-categories is a FinBERT model fine-tuned on about 14,000 manually annotated sentences from firms' ESG reports and annual reports.
finbert-esg-9-categories classifies a text into nine fine-grained ESG topics:
Climate Change, Natural Capital, Pollution & Waste, Human Capital, Product Liability, Community Relations, Corporate Governance, Business Ethics & Values, and Non-ESG. This model complements
finbert-esg which classifies a text into four coarse-grained ESG themes (
E, S, G or None).
Detailed description of the nine fine-grained ESG topic definition, some examples for each topic, training sample, and the model’s performance can be found
here.
Input: A text.
Output: Climate Change, Natural Capital, Pollution & Waste, Human Capital, Product Liability, Community Relations, Corporate Governance, Business Ethics & Values, or Non-ESG.
How to use
You can use this model with Transformers pipeline for fine-grained ESG 9 categories classification.
1from transformers import BertTokenizer, BertForSequenceClassification, pipeline
2
3finbert = BertForSequenceClassification.from_pretrained('yiyanghkust/finbert-esg-9-categories',num_labels=9)
4tokenizer = BertTokenizer.from_pretrained('yiyanghkust/finbert-esg-9-categories')
5nlp = pipeline("text-classification", model=finbert, tokenizer=tokenizer)
6
7results = nlp('For 2002, our total net emissions were approximately 60 million metric tons of CO2 equivalents for all businesses
8 and operations we have financial interests in, based on its equity share in those businesses and operations.')
9print(results) # [{'label': 'Climate Change', 'score': 0.9955655932426453}]
If you use the model in your academic work, please cite the following paper:
Huang, Allen H., Hui Wang, and Yi Yang. "FinBERT: A Large Language Model for Extracting Information from Financial Text." Contemporary Accounting Research (2022).
