A collection of homeworks of market microstructure models.

Baruch MFE 2019 Spring

No description available

No description available

C++ framework for simulating high-frequency trading (HFT) strategies, order book mechanics, and market microstructure, ideal for developing, testing, and analyzing algorithmic trading models in a realistic setting.

End-to-end ML pipeline for high-frequency market microstructure analysis. Engineers features from order book and trade data, fits Hawkes processes, detects market regimes via HMM, and builds tree-based models for mid-price prediction.

R package for market microstructure models and data

A production-ready AI-powered trading system achieving 78% directional accuracy using advanced deep learning models for real-time market microstructure analysis.

Implementation of market risk management models: Value-at-Risk (VaR) estimation, analysis of transaction impact in market microstructure, and option pricing simulation using Black-Scholes and Monte Carlo methods

Full-stack cryptocurrency price prediction system leveraging CNN, LSTM, and CNN-LSTM models. Features algorithmic trading, multi-timeframe analysis, neural networks, and high-frequency market microstructure insights.

Code, models, and notes for FE570 - Market Microstructure and Trading Strategies.

A market microstructure simulator with limit order books, order flow models, and agent-based trading strategies.

This repository is for the laboratory #2 of the subject Microstructure and Trading Systems. With the objective to apply the APT and Roll Models to the market microstructure.

A production-ready AI-powered trading system achieving 78% directional accuracy using advanced deep learning models for real-time market microstructure analysis.

End to end market microstructure pipeline for processing Level 2 order book data and evaluating order flow signals with regime aware, distributional models.

Python library for Indian market microstructure simulation, limit order book preprocessing, feature engineering, and event-driven backtesting. Includes realistic execution, market-making and momentum strategies, latency/slippage models, and robustness analysis utilities.

Advanced topics include time series analysis, machine learning in trading, event-based execution, market making, equities and ETF trading-based strategies, market microstructure, event-driven backtesting, and trading with mixture models among others.

Developed a machine learning pipeline to forecast next-week returns of U.S. ESG ETFs using technical, cross-asset and market microstructure features with LightGBM, XGBoost and Random Forest models.

Deterministic tick-level algorithmic research environment for FX markets. Built on pure Python stdlib, integer-only arithmetic, and reproducible simulation. Phase 1+2 complete: time engine, tick data pipeline, microstructure models.

This project models **Level-2 order book dynamics** to forecast short-term market movements and evaluate the profitability of high-frequency trading (HFT) strategies. It combines machine learning (Scikit-learn) with backtesting to analyze market microstructure, transaction costs, and execution risks.

A Next.js web application for analyzing limit order book (LOB) data and market microstructure. This interactive dashboard provides real-time visualization of order book dynamics, volatility analysis, and price prediction using time-series models.

AI trading bot for Kalshi prediction markets. Uses Claude AI + ensemble models to detect mispricings, Kelly criterion position sizing, spread-adjusted stop-loss, sportsbook odds integration, and exploits longshot bias from microstructure research. Live + paper trading modes.

Building a Python-based limit order book that models market microstructure by handling orders, cancellations, and trade execution. Designed a simulation layer to generate random trading activity and implemented analytics to track bid-ask spread, trade prices, and market depth. Extended core matching engine in C++ for performance benchmarking.

This project models human attention as a market microstructure system. The core idea is to treat attention like a tradable asset, where boredom and fatigue act as drift forces, notifications and high-impact posts behave like aggressive orders, and the platform algorithm acts as an adaptive market maker optimizing for long-term engagement.