Week 11: Natural Language Processing in Finance

Resources

• Core Reading
  • Textual Analysis in Finance
• Implementation Examples
  • SEC Filings Analysis: 10-K Item 7
  • SEC Filings: Download and Process
  • Sentiment Analysis Implementation
• Modern NLP Approaches
  • OpenAI API for Text Analysis
  • FinBERT and Financial NLP Models
  • Example Code:

    
    import openai
    from transformers import pipeline
    
    # OpenAI approach
    def analyze_sentiment_openai(text):
        response = openai.ChatCompletion.create(
            model="gpt-3.5-turbo",
            messages=[
                {"role": "system", "content": "Analyze the sentiment of this financial text. Return POSITIVE, NEGATIVE, or NEUTRAL."},
                {"role": "user", "content": text}
            ]
        )
        return response.choices[0].message['content']
    
    # FinBERT approach
    sentiment_analyzer = pipeline("sentiment-analysis", 
                                model="ProsusAI/finbert")
    
    def analyze_sentiment_finbert(text):
        result = sentiment_analyzer(text)
        return result[0]['label']
    
    # Example usage
    text = "Company X reported strong earnings growth and increased market share."
    print(analyze_sentiment_openai(text))    # Requires API key
    print(analyze_sentiment_finbert(text))   # Local model
                                

Assignment

SEC Filings Analysis Project

• Process and analyze 10-K SEC filings
• Compare traditional and LLM-based sentiment analysis