alasteirho/FIN-RoBERTa-Custom

Financial Sentiment Analysis RoBERTa

A fine-tuned RoBERTa model for financial sentiment analysis that classifies text into negative, neutral, or positive sentiment.

Model Description

This model is a fine-tuned version of roberta-base specifically trained for financial text sentiment classification. It achieves 98.72% accuracy and 98.44% F1 (macro) on the Financial PhraseBank (100% agreement) benchmark.

Labels

Label	ID	Description
negative	0	Negative financial sentiment
neutral	1	Neutral financial sentiment
positive	2	Positive financial sentiment

Usage

from transformers import RobertaForSequenceClassification, RobertaTokenizer
import torch

# Load model and tokenizer
model_name = "alasteirho/alasteirho/FIN-RoBERTa-Custom" 
tokenizer = RobertaTokenizer.from_pretrained(model_name)
model = RobertaForSequenceClassification.from_pretrained(model_name)
model.eval()

# Predict sentiment
def predict_sentiment(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=128)
    with torch.no_grad():
        outputs = model(**inputs)
        probs = torch.softmax(outputs.logits, dim=1)
        pred = torch.argmax(probs, dim=1).item()
    
    labels = ["negative", "neutral", "positive"]
    return labels[pred], probs[0].tolist()

# Example
text = "The company reported strong quarterly earnings, beating analyst expectations."
sentiment, confidence = predict_sentiment(text)
print(f"Sentiment: {sentiment}")
print(f"Confidence: {confidence}")

Using Pipeline (Simple)

from transformers import pipeline

classifier = pipeline("sentiment-analysis", model="alasteirho/FIN-RoBERTa-Custom")
result = classifier("Revenue declined 15% year-over-year due to weak demand.")
print(result)

Training Data

The model was trained on a combined dataset from multiple financial sentiment sources:

Dataset	Samples	Description
Financial PhraseBank (50% agree)	~4,840	Financial news sentences
Twitter Financial News	~9,500	Financial tweets
FiQA 2018	~1,100	Financial QA sentiment
SemEval 2017 Task 5	~1,100	Financial news headlines

Total training samples: ~14,000 (after deduplication)

Evaluation Results

Financial PhraseBank (100% Agreement) - Held-out Test Set

Metric	Score
Accuracy	0.9872
F1 (Macro)	0.9844
F1 (Weighted)	0.9872

Per-Class Performance

Class	Precision	Recall	F1-Score	Support
Negative	0.97	0.99	0.98	303
Neutral	0.99	0.99	0.99	1,391
Positive	0.98	0.98	0.98	570

Confusion Matrix

              Predicted
           neg  neu  pos
Actual neg  300    1    2
       neu    5 1376   10
       pos    3    8  559

Training Procedure

Hyperparameters

Parameter	Value
Base Model	roberta-base
Epochs	2 (early stopping)
Batch Size	16
Gradient Accumulation	2
Effective Batch Size	32
Learning Rate	2e-5
Weight Decay	0.01
Warmup Ratio	0.1
LR Scheduler	Linear
Max Sequence Length	128
FP16	True

Training Metrics (Best Checkpoint - Epoch 2)

• Validation Loss: 0.3658
• Validation Accuracy: 85.93%
• Validation F1 (Macro): 84.55%

Intended Use

This model is designed for:

• Financial news sentiment analysis
• Stock/market sentiment classification
• Financial social media sentiment analysis
• Earnings call/report sentiment extraction

Limitations

• Trained primarily on English financial text
• May not generalise well to non-financial domains
• Performance may vary on very long documents (max 128 tokens)
• Cryptocurrency and emerging market terminology may be underrepresented

License

MIT License