roberta-large-financial-news-sentiment-en Open-source Model - Free Analysis of Canadian Financial News Sentiment

Roberta Large Financial News Sentiment En

Developed by Jean-Baptiste

This model is fine-tuned for sentiment classification of financial news (especially Canadian news), trained on a mixed dataset, and particularly suitable for Canadian financial news analysis.

Text Classification

Transformers

EnglishOpen Source License:MIT #Financial Sentiment Analysis #Canadian Market Optimization #High Precision F1-93%

Downloads 969

Release Time : 12/28/2022

Model Overview

A financial news sentiment classification model fine-tuned based on the RoBERTa-large architecture, specifically designed to analyze sentiment tendencies (negative/neutral/positive) in financial news texts, with excellent performance on Canadian financial news.

Model Features

Specialized Optimization for Canadian Financial News

Additional training on 2000 annotated Canadian financial news articles, achieving an F1 score of 83.6% in this domain

High-Quality Annotated Data

Only sentences with at least 75% annotator agreement were retained to ensure label reliability

Three-Class Fine-Grained Classification

Distinguishes between negative/neutral/positive sentiment states, rather than simple binary classification

Model Capabilities

Financial text sentiment analysis

News sentiment classification

Specialized analysis for Canadian market news

Use Cases

Financial Market Analysis

Sentiment Monitoring of Corporate Earnings Reports

Analyzing sentiment tendencies in corporate earnings news

Can identify 'revenue growth of 17%' as positive and 'net income decline of 3%' as negative

Market Risk Warning

Detecting negative news events such as bankruptcy announcements

Accurately identifies 'filing for bankruptcy protection' as negative sentiment (confidence >93%)

Investment Decision Support

Analysis of Mining Company Production Reports

Assessing sentiment tendencies in mining company production announcements

Correctly classifies 'solid production performance' as positive sentiment

🚀 Financial News Sentiment Classification Model

This model is fine - tuned from roberta - large for sentiment classification of financial news, with an emphasis on Canadian news.

🚀 Quick Start

This model is designed for sentiment classification of financial news, especially focusing on Canadian news. It was trained on the financial_news_sentiment_mixte_with_phrasebank_75 dataset, which is a customized version of the phrasebank dataset. Only sentences validated by at least 75% of annotators were kept, and approximately 2000 manually - validated Canadian financial news articles were added. The model achieves an overall f1 score of 93.25% and 83.6% on Canadian news.

✨ Features

Customized Dataset: Based on a customized version of the phrasebank dataset with additional Canadian financial news articles.
High Performance: Achieves high f1 scores both overall and on Canadian news.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

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

from transformers import AutoTokenizer, AutoModelForSequenceClassification

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

Advanced Usage

Process a 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': 'negative', 'score': 0.9399105906486511}]

📚 Documentation

Training data

The training data was classified as follows:

Property	Details
Model Type	Fine - tuned from roberta - large for financial news sentiment classification
Training Data	Classified into three classes: 0 (negative), 1 (neutral), 2 (positive)

Model performances

Overall f1 score (average macro)

precision	recall	f1
0.9355	0.9299	0.9325

By entity

entity	precision	recall	f1
negative	0.9605	0.9240	0.9419
neutral	0.9538	0.9459	0.9498
positive	0.8922	0.9200	0.9059

📄 License

This project is licensed under the MIT license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご