Tipo di strategia
Algoritmi di trading con machine learning
Modelli ML open source per la finanza quantitativa
Gli algoritmi di trading con machine learning applicano modelli di apprendimento statistico (gradient boosting, reti neurali, Transformer) a serie temporali finanziarie per prevedere i movimenti dei prezzi o generare segnali di trading. Queste implementazioni open source coprono sia formulazioni di regressione (target di prezzo) sia di classificazione (direzione).
12 algoritmi2 librerie
Rete di algoritmi
Come gli algoritmi ML si collegano tra le librerie
ML Algorithms
Freqtrade9 algos
Microsoft Qlib6 algos
LightGBMRegressorintermediate
LightGBMClassifierintermediate
XGBoostRegressorintermediate
XGBoostClassifierintermediate
PyTorchMLPRegressoradvanced
PyTorchTransformerRegressoradvanced
LGBModelintermediate
XGBModelintermediate
DNNModeladvanced
ALSTMadvanced
TFTModeladvanced
GATsadvanced
Pipeline di decisione di trading
Come gli algoritmi ML cooperano in un sistema di trading
1
Market Data Input
Price, volume, indicators
OHLCV price data
Technical indicators
Feature engineering
2
ML Signal Generation
Model predictions
LightGBM/XGBoost regression
Transformer sequence modeling
MLP classification
3
Entry Decision
Signal threshold filtering
Confidence score > 0.6
Direction: Long/Short
4
Exit Strategy
Profit taking & stop loss
Take profit (fixed % or trailing)
Stop loss (ATR-based)
5
Risk Management
Position sizing & portfolio
Position sizing (Kelly criterion)
Portfolio rebalancing
Confronto degli algoritmi
Confronta gli algoritmi ML su dimensioni chiave
Matrice di confronto degli algoritmiFai clic su una colonna per espandere i dettagli
| Metrica | LightGBMRegressorFreqtrade | LightGBMClassifierFreqtrade | XGBoostRegressorFreqtrade | XGBoostClassifierFreqtrade | PyTorchMLPRegressorFreqtrade | PyTorchTransformerRegressorFreqtrade | LGBModelQlib (Microsoft) | XGBModelQlib (Microsoft) | DNNModelQlib (Microsoft) | ALSTMQlib (Microsoft) | TFTModelQlib (Microsoft) | GATsQlib (Microsoft) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ComplessitΓ | βββintermediate | βββintermediate | βββintermediate | βββintermediate | ββββadvanced | ββββadvanced | βββintermediate | βββintermediate | ββββadvanced | ββββadvanced | ββββadvanced | ββββadvanced |
| Tipo di previsione | Regressione | Classificazione | Regressione | Classificazione | Regressione | Regressione | Misto | Misto | Misto | Sequenza | Misto | Misto |
| VelocitΓ di addestramento | β‘β‘β‘ | β‘β‘β‘ | β‘β‘β‘ | β‘β‘β‘ | β‘β‘ | β‘ | β‘β‘ | β‘β‘ | β‘β‘ | β‘ | β‘β‘ | β‘β‘ |
| Accuratezza | πππ | πππ | πππ | πππ | πππ | ππππ | ππ | ππ | ππ | πππ | ππ | ππππ |
| Ideale per | Dati tabellari | Dati tabellari | Dati tabellari | Dati tabellari | Pattern non lineari | Pattern di serie temporali | Generico | Generico | Generico | Dati sequenziali | Generico | Relazioni a grafo |
Freqtrade
Gradient boosting regression model for price movement prediction using LightGBM.
Sorgente:
freqtrade/freqai/prediction_models/LightGBMRegressor.pyLightGBM classification model for user-defined discrete labels.
Sorgente:
freqtrade/freqai/prediction_models/LightGBMClassifier.pyXGBoost-based regression model for continuous value prediction.
Sorgente:
freqtrade/freqai/prediction_models/XGBoostRegressor.pyXGBoost classification model for user-defined discrete labels.
Sorgente:
freqtrade/freqai/prediction_models/XGBoostClassifier.pyMulti-layer perceptron neural network for regression-based price prediction.
Sorgente:
freqtrade/freqai/prediction_models/PyTorchMLPRegressor.pyQlib (Microsoft)
LightGBM model for stock return prediction using technical and fundamental features.
Parametri chiave
| num_leaves | 31 | Maximum number of leaves |
| learning_rate | 0.1 | Boosting learning rate |
Sorgente:
qlib/contrib/model/gbdt.pyXGBoost model for stock return prediction.
Sorgente:
qlib/contrib/model/xgboost.pyDeep neural network for nonlinear feature extraction and return prediction.
Parametri chiave
| hidden_size | 256 | Hidden layer size |
Sorgente:
qlib/contrib/model/pytorch_nn.pyAttention-based LSTM for sequential stock data modeling with attention mechanism.
Parametri chiave
| hidden_size | 64 | LSTM hidden size |
| num_layers | 2 | Number of LSTM layers |
Sorgente:
qlib/contrib/model/pytorch_alstm.pyTemporal Fusion Transformer combining static and temporal features for multi-horizon prediction.
Sorgente:
qlib/contrib/model/pytorch_tft.pyAlgoritmi di trading con machine learning, riferimento algoritmi
- LightGBMRegressor (Freqtrade)
- Gradient boosting regression model for price movement prediction using LightGBM. Sorgente: https://github.com/freqtrade/freqtrade/blob/c0a991f5632cb900075ff2864b6671ed71a319d7/freqtrade/freqai/prediction_models/LightGBMRegressor.py.
- LightGBMClassifier (Freqtrade)
- LightGBM classification model for user-defined discrete labels. Sorgente: https://github.com/freqtrade/freqtrade/blob/c0a991f5632cb900075ff2864b6671ed71a319d7/freqtrade/freqai/prediction_models/LightGBMClassifier.py.
- XGBoostRegressor (Freqtrade)
- XGBoost-based regression model for continuous value prediction. Sorgente: https://github.com/freqtrade/freqtrade/blob/c0a991f5632cb900075ff2864b6671ed71a319d7/freqtrade/freqai/prediction_models/XGBoostRegressor.py.
- XGBoostClassifier (Freqtrade)
- XGBoost classification model for user-defined discrete labels. Sorgente: https://github.com/freqtrade/freqtrade/blob/c0a991f5632cb900075ff2864b6671ed71a319d7/freqtrade/freqai/prediction_models/XGBoostClassifier.py.
- PyTorchMLPRegressor (Freqtrade)
- Multi-layer perceptron neural network for regression-based price prediction. Sorgente: https://github.com/freqtrade/freqtrade/blob/c0a991f5632cb900075ff2864b6671ed71a319d7/freqtrade/freqai/prediction_models/PyTorchMLPRegressor.py.
- PyTorchTransformerRegressor (Freqtrade)
- Transformer architecture for time-series regression using self-attention mechanism. Sorgente: https://github.com/freqtrade/freqtrade/blob/c0a991f5632cb900075ff2864b6671ed71a319d7/freqtrade/freqai/prediction_models/PyTorchTransformerRegressor.py.
- LGBModel (Qlib (Microsoft))
- LightGBM model for stock return prediction using technical and fundamental features. Parametri chiave: num_leaves (Maximum number of leaves), learning_rate (Boosting learning rate).Sorgente: https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/gbdt.py.
- XGBModel (Qlib (Microsoft))
- XGBoost model for stock return prediction. Sorgente: 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. Parametri chiave: hidden_size (Hidden layer size).Sorgente: 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. Parametri chiave: hidden_size (LSTM hidden size), num_layers (Number of LSTM layers).Sorgente: 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. Sorgente: 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. Sorgente: https://github.com/microsoft/qlib/blob/main/qlib/contrib/model/pytorch_gats.py.