v.0.13.3

📢 What’s New

• Simd Accelerated Distributions: link

Full Changelog: v.0.13.1...v.0.13.2

v.0.13.1

📢 What’s New

• Itô Formula Numerical Solver — Now supports both diffusion and jump-diffusion processes.
• General SDE Solver — Added support for Euler–Maruyama, Milstein, SRK2, and SRK4 (Runge–Kutta) schemes. Compatible with both ordinary and fractional noise.

🔗 Full Changelog: v.0.13.0...v.0.13.1

v.0.13.0

🚀 CUDA Support for Fractional Processes

• CUDA Support for Fractional Processes: This release introduces GPU acceleration for fractional stochastic process simulations. The CUDA implementation can significantly improve performance, making large-scale simulations more efficient.

v.0.12.1

New Features:

• AR, MA, ARIMA: Implemented AutoRegressive (AR), Moving Average (MA), and ARIMA models for time series forecasting.
• SARIMA: Seasonal ARIMA model added for seasonal data analysis.
• ARCH, GARCH: Implemented AutoRegressive Conditional Heteroskedasticity (ARCH) and Generalized ARCH (GARCH) models for volatility modeling.
• EGARCH, TGARCH, AGARCH: Added Exponential GARCH (EGARCH), Threshold GARCH (TGARCH), and Asymmetric GARCH (AGARCH) models for advanced volatility modeling.

These models provide robust tools for analyzing and forecasting time series with varying levels of volatility and seasonality.

What's Changed

• feat: add auto regressive processes by @dancixx in #15
• feat: ar models by @dancixx in #16

Full Changelog: v.0.12.0...v.0.12.1

v.0.12.0

🚀 New Features

• Heston Calibration using AI: Improved calibration process leveraging AI for better accuracy.
• Neural Network for Heston Calibration: Enhanced the calibration process with a more effective neural network approach.
• Fractional Ornstein-Uhlenbeck (FOU) Parameter Estimator: Introduced multiple FOU estimators for better modeling flexibility.
• FOU Estimator with Custom Jumps: Added support for jump processes in FOU estimation.
• Fractional Gaussian Noise (FGN) Speed Test: Implemented performance benchmarking for FGN simulation.
• Copula Analysis Module: Added support for Clayton, Gumbel, and empirical Gaussian copulas.
• Kendall’s Tau & Spearman Correlation for Copulas: Introduced statistical tools for dependency analysis.
• Multivariate Copula Interface: in progress...
• Gaussian Kernel Density Estimator (KDE): Added a KDE implementation for density estimation.
• Loss Score for Heston Pricing: Introduced loss-based scoring to improve Heston model pricing.
• Calibration History for Heston & BSM Pricers: Added tracking of past calibration attempts for better diagnostics.

🔧 Fixes & Improvements

• Improved Batching for Heston NN Calibration: Optimized neural network calibration with better data batching.
• Fixed Initial Parameters in Heston Calibration: Addressed issues with incorrect parameter initialization.
• Optional Delta for FOU Estimator: Improved flexibility in parameter estimation.
• Reworked Copula Implementations: Refactored Clayton, Frank, Gumbel, and independence copulas for better accuracy.
• Added Missing Bivariate Implementations: Ensured more complete support for bivariate modeling.
• Updated Plotting Features: Improved visualization for model outputs.
• Reorganized Correlation Module: Better structured for easier use.
• Refactored Calibration Metrics: Enhanced tracking and reporting of calibration errors.
• Reexported iTransformer Interfaces: Improved usability of transformation components.

⚠️ Known Issues

• Heston Calibration Unexpected Behavior: Users may encounter inconsistencies in Heston calibration. Details and discussion can be found #14

Let us know if you encounter any issues or have feature requests! 🚀

Full Changelog: v.0.11.0...v.0.12.0

v.0.11.0

Release Notes

• Heston Model Calibration Updates:

  • Refined Heston calibrator with improvements for stability and accuracy in parameter estimation.
  • Updated calibration process for better integration with volatility surfaces and implied volatility computations.

• Jump Process Enhancements:

  • Reworked various jump processes, including the introduction of Merton jump pricer and KOU model for enhanced modeling capabilities in stochastic environments.

• New Stochastic Processes:

  • Added support for CGMY and tempered stable processes, expanding the library's modeling scope for heavy-tailed and Lévy-driven processes.
  • Introduced the ADG and 2-factor CIR models, providing advanced options for interest rate modeling.

• Mallinavin Derivatives and Isonormal Generator:

  • Implemented improved calculations for Malliavin derivatives, allowing for enhanced sensitivity analysis in stochastic calculus applications.
  • Added isonormal generator, facilitating the generation of Gaussian noise for various processes.

• Option Pricing Extensions:

  • Added finite differences pricer and Asian option pricer, enhancing the library’s range of pricing models.
  • Implemented Black-Scholes implied volatility calculation and integrated it within the Heston pricer for cohesive volatility analysis.

• Versioning and Documentation:

  • Regular version bumps and documentation updates for improved clarity and usability.
  • Organized and refined calibration logic, enhancing code readability and maintainability.

• CI/CD and Code Quality:

  • Updated rust.yml to support continuous integration and testing.
  • Integrated Codecov for improved test coverage monitoring, ensuring consistent code quality.

Full Changelog: v.0.10.0...v.0.11.0

v.0.10.0

What's news?

The stochastic module has been refactored to provide a cleaner and more user-friendly API.

Full Changelog: v.0.9.2...v.0.10.0

v.0.9.2

What's new?

• CEV process was added
• Malliavin derivate of CEV, GBM and vol of SABR
• Docs improvements

What's Changed

• Add license scan report and status by @fossabot in #10

New Contributors

• @fossabot made their first contribution in #10

Full Changelog: v.0.9.0...v.0.9.2

v.0.9.0

Stochastic Updates:

• General Multithreaded Process Generation: Enhanced performance and efficiency in stochastic process simulations.
• Improved FGN-based Models Data Generation: Better data generation from Fractional Gaussian Noise models.
• Hull-White One-Factor Model: Newly implemented for more accurate interest rate modeling.
• Hull-White Two-Factor Model: Added for advanced interest rate predictions.
• Heston 3/2 Model: Integrated into our suite for sophisticated volatility modeling.

Quantitative Finance:

• Heston Pricing Model: Implemented for pricing options under the Heston model assumptions.
• Heston Model Calibration: Tools for calibrating the Heston model to market data.
• Yahoo API module: Tool for retrieve data for testing

Options:

• BSM (Black-Scholes-Merton Model): Enhanced implementation of this fundamental option pricing model.

Bonds:

• CIR (Cox-Ingersoll-Ross Model): Updated with new analytical methods.
• Vasicek Model: Improved stability and performance.
• Hull-White Model: Note that this implementation is currently unstable and should be used with caution.

AI Module:

• FOU-LSTM: A combination of Fractional Ornstein-Uhlenbeck processes with Long Short-Term Memory networks (under construction).
• FOU-VAE: Integration of Fractional Ornstein-Uhlenbeck processes with Variational Autoencoders (under construction).

Statistical Tools:

• Fractal Dimension Analysis: New methods such as Higuchi and variogram approaches added.
• NMLE for Heston Parameters Estimation: Nonlinear Maximum Likelihood Estimation methods implemented for parameter estimation in the Heston model.
• CIR Future Value, PDF, and Asymptotic PDF: New analytical tools for deeper insights into the Cox-Ingersoll-Ross model.

Breaking Changes:

• Finalized Library Structure and Naming Conventions: This release includes breaking changes to the library structure and naming conventions. Please see the migration guide attached for instructions on updating your existing codebase to work with the new version.

stochastic-rs v.0.8.2

Bug fixes 🐞

• Fix: Cir based models to avoid negative values
• Fix: Correlated models rho condition
• Fix: Correlated Cgns return value

Full Changelog: v.0.8.0...v.0.8.2