IA y mercados

Machine learning, NLP, datos alternativos y riesgos de overfitting en finanzas.

En esta página

Data, models and new computation in markets.

This is not “AI that predicts the market”. These are tools for analysis, features, risk, execution and research — with real risks: overfitting, black boxes, data leakage, fragility and governance.


Para qué sirve

This discipline studies how large data sets, machine learning and advanced computation (including quantum in research phase) support analysis, forecasting, risk management, execution and strategy building.

Area Applications Limits
Classical ML Features, classification, regression Overfitting, regime change
Deep learning / LLM News NLP, agents, summarisation Black box, hallucination
Alternative data Satellite, card, web, social Cost, quality, survivorship
Quantum (research) Portfolio opt, pricing, Monte Carlo Immature hardware, hybrid classical

In quantum finance the most studied areas are portfolio optimisation, derivatives pricing, Monte Carlo simulations and quantum ML — today often in hybrid workflows, not immediate replacement of classical systems.


Core concepts

Término In Cyclepedia
AI / ML / deep learning
Big data / alternative data
NLP / sentiment / news analytics
Feature engineering / selection
Train / test / validation set
Cross-validation / data leakage
Overfitting / black box / XAI
Regime detection / anomaly detection
Quantum computing / QML
Post-quantum security

Infrastructure: Matching engine · Execution latency · Market Maker


Tool Role
Data pipeline (tick → feature) ML preparation
ML backtest with purge/embargo Avoid leakage
LLM + RAG on financial documents Assisted research
Agentic workflows Analysis automation (with supervision)
Quantum simulators (Qiskit, etc.) QML prototypes
MLOps / drift monitoring Model production

Researchers, desks and institutions

Figure / entity Contribution
Jim Simons Renaissance, statistics at scale
Marcos López de Prado ML in finance, validation
D.E. Shaw / David Shaw Computational finance
Two Sigma / AQR Quant + alternative data
Doyne Farmer / J.-P. Bouchaud Complexity and microstructure
IBM Quantum / academic research Quantum finance prototypes
Patrick Rebentrost, Seth Lloyd, Marco Pistoia Quantum algorithms for finance

  1. Quantitative trading — quant foundation
  2. Backtest · Sample size — data and samples
  3. Systematic trading — models and metrics
  4. Overfitting · Out of sample — ML backtest and leakage
  5. Forward test · Paper trading — validation
  6. Market regime — context and drift
  7. Anti-scam — risks, fake bots, unrealistic promises

Prerequisite: Systematic trading.


Enlaces