🚀 AI Market Trend Analysis Project

A beginner-friendly machine learning project that predicts stock market trends using technical indicators and advanced ML algorithms.

📄 Project Report: View Report
📊 Live Dashboard: Launch Dashboard
🎥 Video Demo & Presentation: Watch Demo

📋 Table of Contents

• Overview
• Features
• Quick Start
• Project Structure
• Installation
• Model Performance
• Technical Details
• Disclaimer

🎯 Overview

This project demonstrates how to build an end-to-end AI system for stock market trend prediction. It's designed for beginners but includes advanced techniques that make it suitable for learning and real-world applications.

What it does:

• Collects real-time stock data from Yahoo Finance
• Engineers 40+ technical indicators (RSI, MACD, Bollinger Bands, etc.)
• Trains multiple ML models (Random Forest, XGBoost, Logistic Regression)
• Provides an interactive web dashboard for predictions
• Achieves 60%+ accuracy on market trend prediction

Target Audience: Students, beginners in ML/Finance, and anyone interested in algorithmic trading

✨ Features

🔄 Data Collection

• Real-time data from Yahoo Finance API
• Multiple stocks support (AAPL, GOOGL, MSFT, AMZN, TSLA)
• Flexible time periods (1Y, 2Y, 5Y, Max)
• Automatic data cleaning and validation

🔧 Feature Engineering

• 40+ Technical Indicators:
  • Moving Averages (SMA, EMA)
  • Momentum (RSI, ROC, Stochastic)
  • Volatility (Bollinger Bands, ATR)
  • Volume (OBV, VPT, PVT)
  • Price Action patterns

🤖 Machine Learning

• Multiple Models: Random Forest, XGBoost, Logistic Regression
• Automatic Model Selection: Best model chosen by cross-validation
• Class Balancing: Handles imbalanced market data
• Feature Importance: Understand which indicators matter most

📊 Interactive Dashboard

• Real-time Predictions: Live market trend forecasting
• Beautiful Visualizations: Interactive charts with Plotly
• Model Insights: Feature importance and performance metrics
• User-friendly Interface: Built with Streamlit

🚀 Quick Start

1. Clone & Setup

git clone https://github.com/23f2000792/ai-market-trend-analysis
cd ai-market-trend-analysis
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt

2. Collect Data (Week 1)

python data_collector.py

This downloads 5 years of stock data for major tech companies.

3. Engineer Features (Week 2)

python feature_engineer.py

Creates 40+ technical indicators from raw price data.

4. Train Models (Week 3)

python model_trainer.py

Trains and evaluates multiple ML models, saves the best one.

5. Launch Dashboard (Week 4)

streamlit run streamlit_app/app.py

Opens interactive web dashboard at http://localhost:8501

📁 Project Structure

ai-market-trend-analysis/
│
├── 📂 data/
│   ├── raw/                 # Raw stock data (CSV files)
│   └── features/            # Processed data with technical indicators
│
├── 📂 models/
│   ├── *.pkl               # Trained ML models
│   ├── scaler.pkl          # Feature scaler
│   ├── model_metadata.json # Model performance info
│   └── feature_importance.csv
│
├── 📂 notebooks/           # Jupyter notebooks for exploration
│   └── 01_data_exploration.ipynb
│
├── 📂 streamlit_app/       # Interactive dashboard
│   └── app.py
│
├── 📂 src/                 # Core modules
│   ├── data_collector.py   # Stock data collection
│   ├── feature_engineer.py # Technical indicators
│   └── model_trainer.py    # ML model training
│
├── 📋 requirements.txt     # Python dependencies
├── 📚 README.md           # This file
└── 🚀 setup.py            # Installation script

🛠 Installation

Prerequisites

• Python 3.8+ (recommended: 3.9 or 3.10)
• Git
• Internet connection (for data fetching)

Step-by-Step Installation

1. Clone the repository

git clone https://github.com/23f2000792/ai-market-trend-analysis
cd ai-market-trend-analysis

2. Create virtual environment

python -m venv venv

3. Activate virtual environment

# On Windows
venv\Scripts\activate

# On macOS/Linux
source venv/bin/activate

4. Install dependencies

pip install -r requirements.txt

📊 Model Performance

Default Performance (5-year data, 5 stocks)

Model	Accuracy	F1-Score	Training Time	Notes
Random Forest	64.2%	0.639	2.3s	Most balanced performance
XGBoost	66.1%	0.658	5.7s	Highest accuracy
Logistic Regression	61.8%	0.612	0.8s	Fastest training

Key Metrics

• Precision: 62-68% across all classes
• Recall: 58-71% (varies by market condition)
• Cross-validation: 5-fold CV with stratified sampling
• Class Balance: Down(35%), Stable(30%), Up(35%)

Feature Importance (Top 10)

1. RSI_14 (0.124) - Momentum indicator
2. SMA_Ratio_5_20 (0.089) - Trend strength
3. MACD_Histogram (0.076) - Momentum divergence
4. BB_Position (0.071) - Volatility position
5. Volume_Ratio (0.068) - Volume confirmation
6. Price_to_SMA50 (0.063) - Long-term trend
7. ATR (0.059) - Volatility measure
8. Stoch_K (0.057) - Overbought/oversold
9. ROC_10 (0.054) - Price momentum
10. EMA_12 (0.051) - Short-term trend

🔬 Technical Details

Data Pipeline

1. Collection: Yahoo Finance API → Raw OHLCV data
2. Cleaning: Handle missing values, outliers, stock splits
3. Feature Engineering: Create 40+ technical indicators
4. Preprocessing: Scale features, encode targets
5. Training: Multiple models with cross-validation
6. Evaluation: Comprehensive metrics and validation

Technical Indicators Implemented

Trend Indicators

• Simple Moving Average (SMA): 5, 10, 20, 50, 200 periods
• Exponential Moving Average (EMA): 12, 26, 50 periods
• Moving Average Ratios and Cross-overs

Momentum Indicators

• Relative Strength Index (RSI): 14-period
• Rate of Change (ROC): 5, 10, 20 periods
• MACD: Standard 12-26-9 configuration
• Stochastic Oscillator: %K and %D

Volatility Indicators

• Bollinger Bands: 20-period with 2 standard deviations
• Average True Range (ATR): 14-period
• Historical Volatility: 10 and 30-day annualized

Volume Indicators

• On Balance Volume (OBV)
• Volume Price Trend (VPT)
• Price Volume Trend (PVT)
• Volume Moving Averages and Ratios

Model Architecture

Random Forest (Default Best)

RandomForestClassifier(
    n_estimators=100,
    max_depth=10,
    min_samples_split=5,
    min_samples_leaf=2,
    random_state=42
)

XGBoost (Highest Accuracy)

XGBClassifier(
    n_estimators=100,
    max_depth=6,
    learning_rate=0.1,
    random_state=42
)

📚 Learning Resources

APIs and Data Sources

• Yahoo Finance API
• Alpha Vantage
• Quandl

🚨 Disclaimer

IMPORTANT: This project is for educational purposes only.

• 📚 Educational Tool: Designed for learning ML and financial analysis
• ❌ Not Financial Advice: Do not use for actual trading decisions
• 📊 Historical Performance: Past results don't guarantee future performance
• 🎯 Accuracy Limitations: 60-70% accuracy is good for education, not trading
• 💰 Risk Warning: Financial markets involve substantial risk of loss
• 👨‍💼 Professional Advice: Consult qualified financial advisors for investment decisions

Legal Notice

The authors and contributors are not responsible for any financial losses incurred from using this software. Always perform your own research and risk assessment before making investment decisions.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Yahoo Finance for free financial data API
• Streamlit for making beautiful web apps simple
• scikit-learn for excellent ML library
• Plotly for interactive visualizations
• pandas-ta for technical analysis indicators

📞 Support

• 📧 Contact: raushanmahto999@gmail.com

---