Yield Anomaly Detector

Overview

The Yield Anomaly Detector is a highly advanced quantitative analysis tool designed to identify statistical anomalies in financial markets. It replaces basic moving averages and subjective technical analysis with rigorous stochastic mathematics, utilizing the Hurst Exponent and the Ornstein-Uhlenbeck (OU) Stochastic Process.

The system features a Flask-based backend for real-time tick-data processing and a React-style dashboard for visual quant strategy mapping.

Methodology

The core strategy relies on the statistical properties of asset returns and regimes:

1. Hurst Exponent Regime Detection ($H$): Calculates the rolling Hurst Exponent to mathematically define the market regime.
   • $H < 0.45$: Mean-Reverting Regime (Trades enabled)
   • $0.45 \le H \le 0.55$: Random Walk
   • $H > 0.55$: Trending Regime (Mean-reverting trades disabled)
2. Ornstein-Uhlenbeck Process: Calibrates the exact SDE $dX = \theta(\mu - X)dt + \sigma dW$ against real-time data to find the absolute stochastic equilibrium.
3. Anomaly Detection: Signals are generated when $H < 0.45$ and the price statically diverges from the OU Mean ($\mu \pm 2\sigma$).
4. Optimal Exits: Take-profit targets are calculated based on the OU Mean equilibrium $\mu$, maximizing mathematically proven win-rates.

Features

• Real-time Analysis: Polling 1-minute to 15-minute interval data in real time via Yahoo Finance.
• Web Dashboard: Visualizes the OU Model, Hurst Exponent, Volatility bands, and exact Entry/TP/SL limits.
• Discord Integration: Sends alerts to configured webhooks upon confirming actionable signals in the underlying price action.

Installation & Setup

Prerequisites

• Python 3.10 or higher
• Git
• Internet connection (Must allow traffic to fc.yahoo.com for yfinance data polling)

1. Clone the repository:

   git clone https://github.com/Crazyexs/Yield-Anomaly.git
   cd yield-anomaly

Linux / macOS Setup

2. Set up the Virtual Environment & Install Dependencies:

   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt
   
   # Ensure advanced math libraries are installed
   pip install statsmodels hurst yfinance pandas flask requests
   
   # Install TradingView Datafeed (Required for true broker-matching prices)
   pip install git+https://github.com/rongardF/tvdatafeed.git

Windows Setup

2. Set up the Virtual Environment & Install Dependencies:

   python -m venv venv
   venv\Scripts\activate
   pip install -r requirements.txt
   
   # Ensure advanced math libraries are installed
   pip install statsmodels hurst yfinance pandas flask requests
   
   # Install TradingView Datafeed (Required for true broker-matching prices)
   pip install git+https://github.com/rongardF/tvdatafeed.git

Usage

Web Dashboard

Start the application server:

python server.py

Access the dashboard at http://localhost:5001.

CLI Mode

Run the quantitative engine directly in the terminal for a text-based report:

python quant_engine.py

Configuration

Key parameters can be adjusted in server.py and quant_engine.py or via the Dashboard API:

• window: Rolling window size for OU calibration (default: 40).
• ou_threshold: Number of OU standard deviations ($\sigma$) required to trigger anomaly (default: 2.0).
• risk_percent: Risk per trade for position sizing defaults (default: 1.0%).
• interval: Tick frequency 15m (Must respect Yahoo Finance API limits).

Disclaimer

This software is for educational and research purposes only. It does not constitute financial advice. Trading financial markets involves significant risk.