ML4T Platform

🇹🇭 ภาษาไทย

ml4trading.io — ระบบ integrated learning platform สำหรับ machine learning ใน algorithmic trading สร้างโดย Stefan Jansen ประกอบด้วยหนังสือ, case studies, Python libraries, primer topics, agent skills, และ agent lab

หนังสือ 3rd Edition กำลังจะออก June 2026

ภาพรวม Platform

Component	จำนวน	เนื้อหา
Chapters	27	ครอบคลุม 6 ส่วน ตั้งแต่ foundations ถึง production
Case Studies	9	End-to-end strategies (equities, ETFs, crypto, options, futures, forex, commodities)
Python Libraries	5	Production packages ครอบคลุม full workflow
Primer Topics	61	Foundational concepts ใน ML, statistics, quantitative finance
Agent Skills	56	Autonomous workflow tasks with lookahead/leakage/multiple-testing guardrails
Agent Lab	-	AI research environment สำหรับ forecasting และ market insights

ML4T Workflow — 6 ส่วน (3rd Edition)

1. Foundation   → Data & Strategy Setup     (Ch 1-6)
2. Features     → Feature Engineering       (Ch 7-10)
3. Models       → ML Pipeline to Synthesis  (Ch 11-15)
4. Strategy     → Backtest to Execution     (Ch 16-20)
5. Advanced AI  → RL, RAG & Agents          (Ch 21-24)
6. Production   → Deploy & Monitor          (Ch 25-27)

5 Python Libraries

Library	หน้าที่
ML4T Data	Unified market data acquisition จาก 19+ providers
ML4T Engineer	Features, labels, alternative bars, leakage-safe dataset preparation
ML4T Diagnostic	Feature validation, strategy diagnostics, Deflated Sharpe Ratio
ML4T Backtest	Event-driven backtesting with realistic execution
ML4T Live	Production trading with broker integrations

Agent Skills — Design Philosophy

56 skills มี built-in guardrails ป้องกัน: Lookahead bias, Data leakage, Multiple testing errors

3rd Edition — Chapters ทั้งหมด 27 chapters

Ch	ชื่อ	Part
1	The Process Is Your Edge	Foundation
2	The Financial Data Universe	Foundation
3	Market Microstructure	Foundation
4	Fundamental and Alternative Data	Foundation
5	Synthetic Financial Data	Foundation
6	Strategy Research Framework	Foundation
7	Defining the Learning Task	Features
8	Financial Feature Engineering	Features
9	Model-Based Feature Extraction	Features
10	Text Feature Engineering	Features
11	The ML Pipeline (Linear Models)	Models
12	Advanced Models for Tabular Data (GBM)	Models
13	Deep Learning for Time Series	Models
14	Latent Factor Models	Models
15	Causal Machine Learning	Models
16	Strategy Simulation ★NEW	Strategy
17	Portfolio Construction ★NEW	Strategy
18	Transaction Costs	Strategy
19	Risk Management	Strategy
20	Strategy Synthesis	Strategy
21	Reinforcement Learning	Advanced AI
22	RAG for Financial Research ★NEW	Advanced AI
23	Knowledge Graphs ★NEW	Advanced AI
24	Autonomous Agents ★NEW	Advanced AI
25	Live Trading Systems ★NEW	Production
26	MLOps and Governance ★NEW	Production
27	The Systematic Edge	Production

ดูรายละเอียดแต่ละ chapter: ML4T Book 3rd Edition

Related

• ML4T Book 3rd Edition — หนังสือ 3rd edition (27 chapters, June 2026)
• ML4T Book 2nd Edition — หนังสือ 2nd edition (858 หน้า, 23 chapters)
• TradingView MCP — connect Claude Code กับ TradingView Desktop
• Algorithmic Trading — domain concept

---

🇬🇧 English

ml4trading.io — an integrated learning platform for machine learning in algorithmic trading, created by Stefan Jansen. Includes a book, case studies, Python libraries, primer topics, agent skills, and an agent lab.

The 3rd Edition book is coming June 2026.

Platform Overview

Component	Count	Content
Chapters	27	Six parts from foundations to production
Case Studies	9	End-to-end strategies across equities, ETFs, crypto, options, futures, forex, and commodities
Python Libraries	5	Production packages covering the full workflow
Primer Topics	61	Foundational concepts in ML, statistics, and quantitative finance
Agent Skills	56	Autonomous workflow tasks with built-in guardrails
Agent Lab	-	AI-powered research environment for forecasting and market insights

ML4T Workflow — 6 Parts (3rd Edition)

1. Foundation   → Data & Strategy Setup     (Ch 1-6)
2. Features     → Feature Engineering       (Ch 7-10)
3. Models       → ML Pipeline to Synthesis  (Ch 11-15)
4. Strategy     → Backtest to Execution     (Ch 16-20)
5. Advanced AI  → RL, RAG & Agents          (Ch 21-24)
6. Production   → Deploy & Monitor          (Ch 25-27)

5 Python Libraries

Library	Purpose
ML4T Data	Unified market data acquisition from 19+ providers
ML4T Engineer	Features, labels, alternative bars, leakage-safe dataset preparation
ML4T Diagnostic	Feature validation, strategy diagnostics, Deflated Sharpe Ratio
ML4T Backtest	Event-driven backtesting with realistic execution
ML4T Live	Production trading with broker integrations

Agent Skills Design Philosophy

56 skills with built-in guardrails against:

• Lookahead bias — preventing use of future data in features
• Data leakage — preventing test data from contaminating training
• Multiple testing errors — controlling the number of hypotheses tested simultaneously

Related

• ML4T Book 3rd Edition — 3rd edition (27 chapters, June 2026) — full chapter-by-chapter detail
• ML4T Book 2nd Edition — 2020 edition (858 pages, 23 chapters)
• TradingView MCP — connect Claude Code to TradingView Desktop
• Algorithmic Trading — domain concept