FLS-RoBERTa

FLS-RoBERTa is a pre-trained NLP model to classify forward-looking statements (FLS) in financial text, including:

• Financial Statements,
• Earnings Announcements,
• Earnings Call Transcripts,
• Corporate Social Responsibility (CSR) Reports,
• Environmental, Social, and Governance (ESG) News,
• Financial News,
• Etc.

FLS-RoBERTa is built by further training and fine-tuning the RoBERTa Large language model using a large corpus of 10-K, 10-Q, 8-K, Earnings Call Transcripts, CSR Reports, ESG News, and Financial News text, labeled at the sentence level as forward-looking or non-forward-looking.

The model gives softmax outputs for two labels: FLS (Forward-Looking Statement) and Non-FLS (Non-Forward-Looking Statement).

How to classify text:

The easiest way to use the model for single predictions is Hugging Face's text classification pipeline, which only needs a couple lines of code as shown in the following example:

  
from transformers import pipeline
fls_classifier = pipeline("text-classification", model="soleimanian/fls-roberta-large")
print(fls_classifier("We expect revenue to grow by approximately 15% over the next fiscal year as we expand into new markets."))
  

I provide an example script via Google Colab. You can load your data to a Google Drive and run the script for free on a Colab.

Citation and contact:

Please cite this paper when you use the model. Feel free to reach out to mohammad.soleimanian@concordia.ca with any questions or feedback you may have.