Type de stratégie

Algorithmes de trading par apprentissage automatique

Modèles ML open source pour la finance quantitative

Les algorithmes de trading par apprentissage automatique appliquent des modèles d'apprentissage statistique (gradient boosting, réseaux neuronaux, Transformers) à des séries temporelles financières pour prédire les mouvements de prix ou générer des signaux de trading. Ces implémentations open source couvrent les formulations en régression (cible de prix) et en classification (direction).

13 algorithmes2 bibliothèques

Réseau d'algorithmes

Comment les algorithmes ML se connectent entre les bibliothèques

🧠ML Algorithms

🤖

Freqtrade9 algos

🔬

Microsoft Qlib6 algos

LightGBMRegressorintermediate

LightGBMClassifierintermediate

XGBoostRegressorintermediate

XGBoostClassifierintermediate

CatboostRegressorintermediate

PyTorchMLPRegressoradvanced

PyTorchTransformerRegressoradvanced

LGBModelintermediate

XGBModelintermediate

DNNModeladvanced

ALSTMadvanced

TFTModeladvanced

GATsadvanced

Pipeline de décision de trading

Comment les algorithmes ML coopèrent dans un système de trading

📊

Market Data Input

Price, volume, indicators

OHLCV price data

Technical indicators

Feature engineering

🧠

ML Signal Generation

Model predictions

LightGBM/XGBoost regression

Transformer sequence modeling

MLP classification

📈

Entry Decision

Signal threshold filtering

Confidence score > 0.6

Direction: Long/Short

📉

Exit Strategy

Profit taking & stop loss

Take profit (fixed % or trailing)

Stop loss (ATR-based)

🛡️

Risk Management

Position sizing & portfolio

Position sizing (Kelly criterion)

Portfolio rebalancing

Comparaison des algorithmes

Comparer les algorithmes ML selon les dimensions clés

Matrice de comparaison des algorithmesCliquez sur une colonne pour développer les détails

Métrique	LightGBMRegressorFreqtrade	LightGBMClassifierFreqtrade	XGBoostRegressorFreqtrade	XGBoostClassifierFreqtrade	CatboostRegressorFreqtrade	PyTorchMLPRegressorFreqtrade	PyTorchTransformerRegressorFreqtrade	LGBModelQlib (Microsoft)	XGBModelQlib (Microsoft)	DNNModelQlib (Microsoft)	ALSTMQlib (Microsoft)	TFTModelQlib (Microsoft)	GATsQlib (Microsoft)
🎯Complexité	⭐⭐⭐intermediate	⭐⭐⭐intermediate	⭐⭐⭐intermediate	⭐⭐⭐intermediate	⭐⭐⭐intermediate	⭐⭐⭐⭐advanced	⭐⭐⭐⭐advanced	⭐⭐⭐intermediate	⭐⭐⭐intermediate	⭐⭐⭐⭐advanced	⭐⭐⭐⭐advanced	⭐⭐⭐⭐advanced	⭐⭐⭐⭐advanced
📈Type de prédiction	Régression	Classification	Régression	Classification	Régression	Régression	Régression	Mixte	Mixte	Mixte	Séquence	Mixte	Mixte
⚡Vitesse d'entraînement	⚡⚡⚡	⚡⚡⚡	⚡⚡⚡	⚡⚡⚡	⚡⚡	⚡⚡	⚡	⚡⚡	⚡⚡	⚡⚡	⚡	⚡⚡	⚡⚡
🎯Précision	📊📊📊	📊📊📊	📊📊📊	📊📊📊	📊📊📊	📊📊📊	📊📊📊📊	📊📊	📊📊	📊📊	📊📊📊	📊📊	📊📊📊📊
💡Idéal pour	Données tabulaires	Données tabulaires	Données tabulaires	Données tabulaires	Usage général	Motifs non linéaires	Motifs de séries temporelles	Usage général	Usage général	Usage général	Données séquentielles	Usage général	Relations graphiques

Freqtrade

LightGBMRegressor

Freqtrade

Machine Learningintermediate

Gradient boosting regression model for price movement prediction using LightGBM.

Vitesse⚡⚡⚡

Précision📊📊📊

Paramètres clés

n_estimators	1000	Number of boosting rounds
learning_rate	0.01	Step size shrinkage

Source :freqai/prediction_models/LightGBMRegressor.py

LightGBMClassifier

Freqtrade

Machine Learningintermediate

Gradient boosting classification model for directional prediction (up/down/neutral).

Vitesse⚡⚡⚡

Précision📊📊📊

Paramètres clés

n_estimators

1000

Number of boosting rounds

Source :freqai/prediction_models/LightGBMClassifier.py

XGBoostRegressor

Freqtrade

Machine Learningintermediate

XGBoost-based regression model for continuous value prediction.

Vitesse⚡⚡⚡

Précision📊📊📊

Paramètres clés

n_estimators	1000	Number of boosting rounds
max_depth	6	Maximum tree depth

Source :freqai/prediction_models/XGBoostRegressor.py

XGBoostClassifier

Freqtrade

Machine Learningintermediate

XGBoost-based classification model for directional prediction.

Vitesse⚡⚡⚡

Précision📊📊📊

Paramètres clés

n_estimators

1000

Number of boosting rounds

Source :freqai/prediction_models/XGBoostClassifier.py

CatboostRegressor

Freqtrade

Machine Learningintermediate

CatBoost gradient boosting model with native categorical feature support.

Vitesse⚡⚡

Précision📊📊📊

Paramètres clés

iterations

1000

Number of boosting iterations

Source :freqai/prediction_models/CatboostRegressor.py

PyTorchMLPRegressor

Freqtrade

Machine Learningadvanced

Multi-layer perceptron neural network for regression-based price prediction.

Vitesse⚡⚡

Précision📊📊📊

Paramètres clés

hidden_dim	128	Hidden layer dimension
dropout_percent	0.2	Dropout rate

Source :freqai/prediction_models/PyTorchMLPRegressor.py

PyTorchTransformerRegressor

Freqtrade

Machine Learningadvanced

Transformer architecture for time-series regression using self-attention mechanism.

Vitesse⚡

Précision📊📊📊📊

Paramètres clés

num_heads	8	Number of attention heads
num_layers	2	Number of transformer layers

Source :freqai/prediction_models/PyTorchTransformerRegressor.py

Qlib (Microsoft)

LGBModel

Qlib (Microsoft)

Machine Learningintermediate

LightGBM model for stock return prediction using technical and fundamental features.

Vitesse⚡⚡

Précision📊📊📊

Paramètres clés

num_leaves	31	Maximum number of leaves
learning_rate	0.1	Boosting learning rate

Source :qlib/contrib/model/gbdt.py

XGBModel

Qlib (Microsoft)

Machine Learningintermediate

XGBoost model for stock return prediction.

Vitesse⚡⚡

Précision📊📊📊

Source :qlib/contrib/model/xgboost.py

DNNModel

Qlib (Microsoft)

Machine Learningadvanced

Deep neural network for nonlinear feature extraction and return prediction.

Vitesse⚡⚡

Précision📊📊📊

Paramètres clés

hidden_size

256

Hidden layer size

Source :qlib/contrib/model/pytorch_nn.py

ALSTM

Qlib (Microsoft)

Machine Learningadvanced

Attention-based LSTM for sequential stock data modeling with attention mechanism.

Vitesse⚡

Précision📊📊📊

Paramètres clés

hidden_size	64	LSTM hidden size
num_layers	2	Number of LSTM layers

Source :qlib/contrib/model/pytorch_alstm.py

TFTModel

Qlib (Microsoft)

Machine Learningadvanced

Temporal Fusion Transformer combining static and temporal features for multi-horizon prediction.

Vitesse⚡⚡

Précision📊📊📊

Source :qlib/contrib/model/pytorch_tft.py

GATs

Qlib (Microsoft)

Machine Learningadvanced

Graph Attention Networks modeling stock relationships for cross-stock prediction.

Vitesse⚡⚡

Précision📊📊📊

Source :qlib/contrib/model/pytorch_gats.py

Algorithmes de trading par apprentissage automatique, référence des algorithmes

LightGBMRegressor (Freqtrade): Gradient boosting regression model for price movement prediction using LightGBM. Paramètres clés : n_estimators (Number of boosting rounds), learning_rate (Step size shrinkage).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/LightGBMRegressor.py.
LightGBMClassifier (Freqtrade): Gradient boosting classification model for directional prediction (up/down/neutral). Paramètres clés : n_estimators (Number of boosting rounds).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/LightGBMClassifier.py.
XGBoostRegressor (Freqtrade): XGBoost-based regression model for continuous value prediction. Paramètres clés : n_estimators (Number of boosting rounds), max_depth (Maximum tree depth).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/XGBoostRegressor.py.
XGBoostClassifier (Freqtrade): XGBoost-based classification model for directional prediction. Paramètres clés : n_estimators (Number of boosting rounds).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/XGBoostClassifier.py.
CatboostRegressor (Freqtrade): CatBoost gradient boosting model with native categorical feature support. Paramètres clés : iterations (Number of boosting iterations).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/CatboostRegressor.py.
PyTorchMLPRegressor (Freqtrade): Multi-layer perceptron neural network for regression-based price prediction. Paramètres clés : hidden_dim (Hidden layer dimension), dropout_percent (Dropout rate).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/PyTorchMLPRegressor.py.
PyTorchTransformerRegressor (Freqtrade): Transformer architecture for time-series regression using self-attention mechanism. Paramètres clés : num_heads (Number of attention heads), num_layers (Number of transformer layers).Source : https://github.com/freqtrade/freqtrade/blob/develop/freqai/prediction_models/PyTorchTransformerRegressor.py.
LGBModel (Qlib (Microsoft)): LightGBM model for stock return prediction using technical and fundamental features. Paramètres clés : num_leaves (Maximum number of leaves), learning_rate (Boosting learning rate).Source : https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/gbdt.py.
XGBModel (Qlib (Microsoft)): XGBoost model for stock return prediction. Source : https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/xgboost.py.
DNNModel (Qlib (Microsoft)): Deep neural network for nonlinear feature extraction and return prediction. Paramètres clés : hidden_size (Hidden layer size).Source : https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_nn.py.
ALSTM (Qlib (Microsoft)): Attention-based LSTM for sequential stock data modeling with attention mechanism. Paramètres clés : hidden_size (LSTM hidden size), num_layers (Number of LSTM layers).Source : https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_alstm.py.
TFTModel (Qlib (Microsoft)): Temporal Fusion Transformer combining static and temporal features for multi-horizon prediction. Source : https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_tft.py.
GATs (Qlib (Microsoft)): Graph Attention Networks modeling stock relationships for cross-stock prediction. Source : https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_gats.py.