Skip to content

README.md

Latest commit

 

History

History
259 lines (203 loc) · 12.8 KB

README.md

File metadata and controls

259 lines (203 loc) · 12.8 KB
Raw

Appendix - Financial Mathematics Theory and Practical Examples

This directory contains a comprehensive guide to all financial mathematics theory used in quant trading, implemented with easy-to-understand example code.

📚 Structure

Appendix/
├── README.md                           # This file
├── Chapter1_Linear_Algebra/           # Linear Algebra: Portfolio & Factors
│   ├── __init__.py
│   ├── portfolio_optimization.py      # Portfolio optimization example
│   ├── pca_factor_analysis.py         # Factor analysis using PCA
│   ├── factor_regression.py           # Multi-factor regression (Fama-French)
│   └── var_vecm_models.py             # Matrix operations in VAR & VECM models
├── Chapter2_Calculus/                  # Analysis & Calculus: Options & Optimization
│   ├── __init__.py
│   ├── gradient_descent_demo.py       # Gradient descent visualization
│   ├── backpropagation_example.py     # Understanding backpropagation algorithm
│   ├── garch_volatility.py            # Calculus principles in GARCH models
│   ├── wavelet_transform.py           # Wavelet Transform
│   ├── ito_lemma.py                   # Ito's Lemma
│   └── bayesian_optimization.py       # Bayesian Optimization
├── Chapter3_Probability_Statistics/    # Probability & Time Series Statistics
│   ├── __init__.py
│   ├── stationarity_analysis.py       # Stationarity testing and understanding
│   ├── arima_modeling.py              # Probabilistic foundations of ARIMA models
│   ├── cointegration_pairs.py         # Cointegration and pair trading
│   ├── copula_dependence.py            # Dependence analysis using Copula
│   └── monte_carlo_simulation.py       # Monte Carlo simulation
├── Chapter4_Bayesian_Filtering/       # Bayesian Statistics & Filtering
│   ├── __init__.py
│   ├── bayesian_inference.py          # Bayesian inference examples
│   ├── kalman_filter_demo.py          # Understanding Kalman Filter
│   └── state_space_models.py           # State-space models
└── utils/                              # Utility functions
    └── __init__.py

🎯 Learning Objectives

Each Chapter aims to:

  1. Understand fundamental mathematical structures: Intuitively explain core concepts of each mathematical field
  2. Financial mathematics specialization: Real examples applied to financial data
  3. Learn through code: Implement formulas in code to clearly understand concepts
  4. Analogies and explanations: Explain complex mathematics through everyday analogies

🚀 Quick Start

1. Package Installation

# Install from project root
pip install -r requirements.txt

2. Running Examples

Each Chapter's examples can be run independently:

# Chapter 1: Linear Algebra
python Chapter1_Linear_Algebra/portfolio_optimization.py

# Chapter 2: Calculus
python Chapter2_Calculus/gradient_descent_demo.py

# Chapter 3: Probability & Statistics
python Chapter3_Probability_Statistics/stationarity_analysis.py

# Chapter 4: Bayesian
python Chapter4_Bayesian_Filtering/kalman_filter_demo.py

📊 Overview of Modern Mathematics and Financial Mathematics

1️⃣ Classification of Modern Mathematics

Modern mathematics is broadly divided into Pure Mathematics and Applied Mathematics.

Category Subcategory Main Research Areas Financial Relevance Notes
Pure Math Algebra Groups, rings, fields, equations ⭐⭐⭐ Linear algebra is essential
Analysis Limits, continuity, derivatives, integrals ⭐⭐⭐⭐⭐ Core of financial mathematics
Geometry Shapes, space, distance ⭐⭐ Used in data visualization
Topology Continuity, connectivity ⭐ Theoretical research level
Number Theory Integers, primes, congruence ⭐ Cryptography (blockchain)
Logic Proofs, set theory ⭐ Foundation of algorithm design
Applied Math Probability & Statistics Uncertainty, distributions, estimation ⭐⭐⭐⭐⭐ Language of finance
Numerical Analysis Approximation, optimization ⭐⭐⭐⭐ Essential in practice
Differential Equations Dynamics, SDE ⭐⭐⭐⭐ Option pricing, GARCH
Optimization Theory Constraints, objective functions ⭐⭐⭐⭐⭐ Core of portfolio theory
Information Theory Entropy, information content ⭐⭐⭐ Model selection (AIC/BIC)
Graph Theory Networks, connectivity ⭐⭐ System risk
Computational Machine Learning Math Gradient descent, backpropagation ⭐⭐⭐⭐⭐ Core of AI trading

Relevance Legend:

  • ⭐⭐⭐⭐⭐ Absolutely essential (pillar of financial mathematics)
  • ⭐⭐⭐⭐ Very important (frequently used in practice)
  • ⭐⭐⭐ Important (essential in specific areas)
  • ⭐⭐ Optional (advanced applications)
  • ⭐ Indirect (theoretical background or special fields)

2️⃣ Financial Mathematics Learning Path

[Essential Foundations]          [Core Applications]              [Advanced Applications]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
1. Linear Algebra          →  Portfolio Optimization        →  PCA, Factor Models
2. Calculus                →  Option Pricing, Gradient Descent →  Ito Calculus, SDE
3. Probability & Statistics →  Time Series, Risk Management →  Bayesian, Copula
4. Numerical Analysis      →  Optimization, Simulation      →  Monte Carlo

📊 Project-Specific Mathematical Matrix

This matrix organizes the mathematical theories actually implemented in this project by field.

Mathematical Field Core Concepts Financial Applications Key Techniques/Algorithms
Linear Algebra Vectors, matrices, eigenvalues Portfolio optimization, factor models Covariance matrix, regression matrix, PCA
Analysis Derivatives, integrals, limits Deep learning, volatility models Gradient descent, backpropagation, GARCH
Probability & Statistics Distributions, estimation, testing Time series, risk management ARIMA, Kalman, VaR, Copula
Numerical Analysis Optimization, approximation Portfolio, simulation scipy.optimize, Monte Carlo
Geometry Space, distance, dimensions Dimensionality reduction, visualization PCA, t-SNE, distance metrics
Information Theory Entropy, information content Model selection, feature selection AIC/BIC, mutual information
Graph Theory Networks, connectivity Asset relationships, risk Correlation networks, MST

📝 Detailed Mathematical Techniques

1. Linear Algebra

  • Covariance Matrix (Σ): Measures correlation and volatility between assets
  • Regression Matrix: β = (X'X)⁻¹X'y (least squares method)
  • VAR/VECM: Multivariate time series matrix operations
  • PCA: Dimensionality reduction through eigenvalue decomposition

2. Analysis

  • Gradient Descent: Search for minimum following ∇f(x)
  • Backpropagation: Calculate gradients using chain rule
  • GARCH: σ²ₜ = α₀ + α₁ε²ₜ₋₁ + β₁σ²ₜ₋₁
  • Ito's Lemma: Core of stochastic differential equations

3. Probability & Statistics

  • ARIMA: Stationarity testing + autoregressive models
  • Kalman Filter: State estimation through Bayesian updates
  • VaR/CVaR: Extreme loss risk measurement
  • Copula: Modeling dependence independent of marginal distributions

4. Numerical Analysis

  • scipy.optimize: SLSQP, L-BFGS optimization
  • Monte Carlo: Probabilistic simulation
  • Bayesian Optimization: Hyperparameter tuning

📖 Chapter Details

Chapter 1: Linear Algebra

Core Analogy: "Cocktail Recipe"

  • Vector: Amount of each ingredient [Gin 30ml, Tonic 90ml, Lime 10ml]
  • Matrix: Compatibility chart between ingredients (covariance matrix)
  • Eigenvalue decomposition: Extract core flavors (PCA)

Key Topics:

  • Portfolio variance calculation: σ² = wᵀΣw
    • This formula calculates "how risky is this cocktail?"
    • Foundation of MVO (Mean-Variance Optimization) and factor models
  • PCA (Principal Component Analysis):
    • Compresses movements of hundreds of stocks into a few key factors like 'market', 'interest rates', 'oil prices'
  • Meaning and calculation of covariance matrices
  • Matrix operations in Fama-French factor models
  • Matrix representation and estimation of VAR & VECM models

Chapter 2: Analysis & Calculus

Core Analogy: "Descending a mountain in fog"

  • Derivative: Speedometer (measuring rate of change) - "How fast is the price/error changing at this moment?"
  • Gradient descent: Feeling the slope with your feet to descend to the lowest valley (minimum error)
  • Backpropagation: Propagating errors backward to assign responsibility

Key Topics:

  • Optimization through gradient descent
  • Deep Learning (Backpropagation):
    • Principle of how LSTM learns. Differentiates prediction errors to assign responsibility and adjust weights
  • Calculus principles in GARCH models
  • Ito's Lemma:
    • Formula for calculating the rate of change of option prices when stock prices jump randomly (Brownian motion)

Chapter 3: Probability & Time Series Statistics

Core Analogy: "Predicting the future from past weather"

  • Stationarity: Spring (property of returning to original position) - fundamental premise of time series analysis
  • Cointegration: Owner and dog (connected by a leash) - seem to move independently but ultimately move together due to long-term equilibrium

Key Topics:

  • Stationarity testing (ADF Test)
  • Probabilistic foundations of ARIMA models
  • ARIMA / GARCH:
    • Statistically estimates patterns of past data (AR), errors (MA), and volatility (GARCH)
  • Cointegration and pair trading
  • Copula:
    • "Panic Room Effect": Models tail dependence where assets that normally move independently all crash together during crises
  • Probability distributions in GARCH models

Chapter 4: Bayesian Statistics & Filtering

Core Analogy: "Narrowing down suspects with new clues"

  • Bayesian inference: Detective investigation - starts with many suspects (prior probability), updates probability of the real culprit (posterior probability) as evidence (data) emerges
  • Kalman Filter: Navigation (combining GPS + speed) - combines noisy GPS signals (observations) with car speed (model) to estimate 'true position'

Key Topics:

  • Bayesian update (prior → posterior probability)
  • State estimation with Kalman Filter
  • State-Space Models:
    • Tracks unobserved 'true market beta'
  • Prophet:
    • Flexibly decomposes trends and seasonality using Bayesian methods to predict the future
  • Bayesian structure of Prophet model

🔗 Connection to Project

Mathematical Field Core Concepts Main Applications Analogy
Linear Algebra Matrix operations, eigenvalues Portfolio optimization, factor models Cocktail recipe
Calculus Gradient descent, backpropagation Deep learning, optimization Descending in fog
Time Series Statistics Stationarity, cointegration Pair trading Owner and dog
Bayesian Kalman Filter, posterior probability State-space models, Prophet Detective investigation

This guide serves as a map that mathematically supports "why the code works that way", rather than simply listing formulas.

📝 Usage

Each example file can be run independently and follows this structure:

  1. Theory explanation: Mathematical concepts explained with analogies
  2. Basic examples: Simple mathematical examples
  3. Financial application: Applied to real financial data
  4. Visualization: Results expressed as graphs

📚 Code Structure

Each example file is self-contained and includes:

  • Theory explanation: Mathematical concepts explained with intuitive analogies
  • Code implementation: Practical Python code demonstrating the concepts
  • Financial applications: Real-world examples using financial data
  • Visualizations: Graphs and plots to illustrate results

The code itself serves as documentation, with comments explaining the mathematical concepts through everyday analogies.

⚠️ Notes

  • All examples are for educational purposes
  • Sufficient verification is required before using in actual investments
  • Data is automatically downloaded via yfinance

🤝 Contributing

This guide aims to explain all financial mathematics in quant trading in an easy-to-understand way. Please suggest improvements or additional examples!