fin-sentiment

fin-sentiment

Financial News Sentiment Analysis using LSTM

fin-sentiment is a Python package that analyzes the sentiment of financial news headlines using an LSTM-based deep learning model. It helps traders, investors, and analysts gauge market sentiment by classifying news as positive, neutral, or negative.

Features

• 📰 Fetch latest financial news for any company
• 🤖 LSTM-based sentiment classification
• 📊 Aggregate sentiment analysis across multiple headlines
• ⚡ Simple, intuitive API
• 🎯 Trained on real financial news data

Installation

pip install fin-sentiment

Quick Start

from fin_sentiment.news_fetcher import fetch_company_news
from fin_sentiment.analyzer import SentimentAnalyzer

# Step 1: Fetch latest news for a company
news = fetch_company_news("TCS", limit=10)

# Step 2: Initialize sentiment analyzer
analyzer = SentimentAnalyzer()

# Step 3: Analyze each news headline
for article in news:
    sentiment = analyzer.analyze(article["title"])
    print(f"{article['title']} → {sentiment}")

# Step 4: Get overall market sentiment
overall = analyzer.overall_sentiment(news)
print(f"Market Mood: {overall}")

# Step 5: Analyze a custom headline
custom_sentiment = analyzer.analyze("Stock prices surged after positive earnings report")
print(custom_sentiment)  # Output: positive

API Reference

Module: fin_sentiment.news_fetcher

fetch_company_news(company_name: str, limit: int = 10) -> list

Fetches the latest news headlines for a specified company from financial sources.

Parameters:

• company_name (str): Name of the company to fetch news for
• limit (int): Maximum number of headlines to retrieve (default: 10)

Returns: List of news articles with titles and metadata

Example:

news = fetch_company_news("Infosys", limit=5)

Module: fin_sentiment.analyzer

SentimentAnalyzer()

Initializes the LSTM-based sentiment analysis model.

Example:

analyzer = SentimentAnalyzer()

.analyze(text: str) -> str

Analyzes the sentiment of a single text string.

Parameters:

• text (str): The headline or text to analyze

Returns: Sentiment label: "positive", "neutral", or "negative"

Example:

sentiment = analyzer.analyze("Company shares hit record high")
print(sentiment)  # Output: positive

.overall_sentiment(news_list: list) -> str

Aggregates sentiment across multiple news headlines to determine overall market mood.

Parameters:

• news_list (list): List of news articles (with "title" key)

Returns: Overall sentiment: "positive", "neutral", or "negative"

Example:

overall = analyzer.overall_sentiment(news)
print(overall)  # Output: neutral

Example Output

TCS shares rise 5% after strong Q3 results → positive
TCS faces data breach allegations → negative
TCS announces dividend payout → positive
Market closes flat amid mixed signals → neutral

Market Mood: neutral

Model Details

Property	Details
Architecture	LSTM (Long Short-Term Memory)
Framework	PyTorch
Dataset	Financial headlines from multiple market sources
Classes	Positive, Neutral, Negative
Output	Softmax probability distribution → final label

Use Cases

• 📈 Trading Signals: Gauge market sentiment before making trades
• 📰 News Monitoring: Track sentiment trends for specific companies
• 🤖 Automated Analysis: Integrate sentiment into trading bots
• 📊 Market Research: Analyze sentiment patterns over time

Requirements

• Python 3.7+
• PyTorch
• NumPy
• Requests (for news fetching)
• Feedparser

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.