Model fine-tuned from roberta-large for topic classification of financial news (emphasis on Canadian news).

Introduction

This model was train on the topic column of financial_news_sentiment_mixte_with_phrasebank_75 dataset. The topic column was generated using a zero-shot classification model on 11 topics. There was no manual reviews on the generated topics and therefore we should expect misclassifications in the dataset, and therefore the trained model might reproduce the same errors.

Training data

Training data was classified as follow:

How to use roberta-large-financial-news-topics-en with HuggingFace

Load roberta-large-financial-news-topics-en and its sub-word tokenizer :

from transformers import AutoTokenizer, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("Jean-Baptiste/roberta-large-financial-news-topics-en")
model = AutoModelForSequenceClassification.from_pretrained("Jean-Baptiste/roberta-large-financial-news-topics-en")


##### Process text sample (from wikipedia)

from transformers import pipeline

pipe = pipeline("text-classification", model=model, tokenizer=tokenizer)
pipe("Melcor REIT (TSX: MR.UN) today announced results for the third quarter ended September 30, 2022. Revenue was stable in the quarter and year-to-date. Net operating income was down 3% in the quarter at $11.61 million due to the timing of operating expenses and inflated costs including utilities like gas/heat and power")

[{'label': 'quaterly financial release', 'score': 0.8829097151756287}]

Model performances

Overall f1 score (average macro)