Machine learning

Machine learning covers the methods where a model learns from data instead of applying dictated rules: classification, anomaly detection, clustering, time-series forecasting. It’s the building block that separates a quantitative approach from a simple rule-based bot. Its weakness is worth remembering: a model that clings too closely to past data gives superb results in testing and disappointing ones live. And none of it predicts markets — it sorts, it ranks, it saves research time. What that looks like in practice, and at what cost: AI trading: the complete guide.