Tous les livres

Professionnel

Applis À propos Coach Connexion Commencer la lecture

Quantitative Finance · Glossaire

Qu'est-ce que « Meta-labelling » ?

Definition 2.3 Machine Learning for Markets · Chapitre 2 — Targets, Labels and Sample Weights

Meta-labelling labels each event of a primary model with 1 if the trade in the primary model’s direction made money (for example, its triple-barrier return is positive) and 0 otherwise, and trains a secondary model on these labels to estimate the probability that the primary call is right; the probability is then used to filter or size the trade (López de Prado, 2018; Joubert, 2022).

Meta-labelling the primary trade: keeping the trades whose estimated probability of success exceeds 0.50 to 0.55 (100% = the primary model alone). Left, the average return per trade in barrier widths; right, the t-statistic with the number of trades adjusted for overlap. Out of sample, last 40% of the dates. Data: ml_labels.meta.
Figure 2.2. Meta-labelling the primary trade: keeping the trades whose estimated probability of success exceeds 0.50 to 0.55 (100% = the primary model alone). Left, the average return per trade in barrier widths; right, the tt-statistic with the number of trades adjusted for overlap. Out of sample, last 40% of the dates. Data: ml_labels.meta.
Lire dans le chapitre →