adelorenzo/stable
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add XGBoostRegressor for freqAI, fix mypy errors

Browse Source
This commit is contained in:
Emre
2022-09-08 14:12:19 +03:00
committed by Robert Caulk
parent 4c9ac6b7c0
commit 1b6410d7d1
6 changed files with 87 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
freqtrade/freqai/prediction_models/BaseClassifierModel.py
View File

@@ -21,7 +21,7 @@ class BaseClassifierModel(IFreqaiModel):
"""
def train(
self, unfiltered_dataframe: DataFrame, pair: str, dk: FreqaiDataKitchen
self, unfiltered_dataframe: DataFrame, pair: str, dk: FreqaiDataKitchen, **kwargs
) -> Any:
"""
Filter the training data and train a model to it. Train makes heavy use of the datakitchen
@@ -68,7 +68,7 @@ class BaseClassifierModel(IFreqaiModel):
return model
def predict(
self, unfiltered_dataframe: DataFrame, dk: FreqaiDataKitchen, first: bool = False
self, dataframe: DataFrame, dk: FreqaiDataKitchen, first: bool = False, **kwargs
) -> Tuple[DataFrame, npt.NDArray[np.int_]]:
"""
Filter the prediction features data and predict with it.
@@ -79,9 +79,9 @@ class BaseClassifierModel(IFreqaiModel):
data (NaNs) or felt uncertain about data (PCA and DI index)
"""
dk.find_features(unfiltered_dataframe)
dk.find_features(dataframe)
filtered_dataframe, _ = dk.filter_features(
unfiltered_dataframe, dk.training_features_list, training_filter=False
dataframe, dk.training_features_list, training_filter=False
)
filtered_dataframe = dk.normalize_data_from_metadata(filtered_dataframe)
dk.data_dictionary["prediction_features"] = filtered_dataframe

8
freqtrade/freqai/prediction_models/BaseRegressionModel.py
View File

@@ -20,7 +20,7 @@ class BaseRegressionModel(IFreqaiModel):
"""
def train(
self, unfiltered_dataframe: DataFrame, pair: str, dk: FreqaiDataKitchen
self, unfiltered_dataframe: DataFrame, pair: str, dk: FreqaiDataKitchen, **kwargs
) -> Any:
"""
Filter the training data and train a model to it. Train makes heavy use of the datakitchen
@@ -67,7 +67,7 @@ class BaseRegressionModel(IFreqaiModel):
return model
def predict(
self, unfiltered_dataframe: DataFrame, dk: FreqaiDataKitchen, first: bool = False
self, dataframe: DataFrame, dk: FreqaiDataKitchen, first: bool = False, **kwargs
) -> Tuple[DataFrame, npt.NDArray[np.int_]]:
"""
Filter the prediction features data and predict with it.
@@ -78,9 +78,9 @@ class BaseRegressionModel(IFreqaiModel):
data (NaNs) or felt uncertain about data (PCA and DI index)
"""
dk.find_features(unfiltered_dataframe)
dk.find_features(dataframe)
filtered_dataframe, _ = dk.filter_features(
unfiltered_dataframe, dk.training_features_list, training_filter=False
dataframe, dk.training_features_list, training_filter=False
)
filtered_dataframe = dk.normalize_data_from_metadata(filtered_dataframe)
dk.data_dictionary["prediction_features"] = filtered_dataframe

2
freqtrade/freqai/prediction_models/BaseTensorFlowModel.py
View File

@@ -17,7 +17,7 @@ class BaseTensorFlowModel(IFreqaiModel):
"""
def train(
self, unfiltered_dataframe: DataFrame, pair: str, dk: FreqaiDataKitchen
self, unfiltered_dataframe: DataFrame, pair: str, dk: FreqaiDataKitchen, **kwargs
) -> Any:
"""
Filter the training data and train a model to it. Train makes heavy use of the datakitchen

46
freqtrade/freqai/prediction_models/XGBoostRegressor.py Normal file
View File

@@ -0,0 +1,46 @@
import logging
from typing import Any, Dict
import xgboost as xgb
from freqtrade.freqai.data_kitchen import FreqaiDataKitchen
from freqtrade.freqai.prediction_models.BaseRegressionModel import BaseRegressionModel
logger = logging.getLogger(__name__)
class XGBoostRegressor(BaseRegressionModel):
"""
User created prediction model. The class needs to override three necessary
functions, predict(), train(), fit(). The class inherits ModelHandler which
has its own DataHandler where data is held, saved, loaded, and managed.
"""
def fit(self, data_dictionary: Dict, dk: FreqaiDataKitchen, **kwargs) -> Any:
"""
User sets up the training and test data to fit their desired model here
:param data_dictionary: the dictionary constructed by DataHandler to hold
all the training and test data/labels.
"""
xgb.set_config(verbosity=2)
xgb.config_context(verbosity=2)
X = data_dictionary["train_features"]
y = data_dictionary["train_labels"]
if self.freqai_info.get("data_split_parameters", {}).get("test_size", 0.1) == 0:
eval_set = None
else:
eval_set = [(data_dictionary["test_features"], data_dictionary["test_labels"])]
sample_weight = data_dictionary["train_weights"]
xgb_model = self.get_init_model(dk.pair)
model = xgb.XGBRegressor(**self.model_training_parameters)
model.fit(X=X, y=y, sample_weight=sample_weight, eval_set=eval_set, xgb_model=xgb_model)
return model