Source code for tensortrade.features.scalers.min_max_normalizer

# Copyright 2019 The TensorTrade Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License

import pandas as pd
import numpy as np

from gym import Space
from copy import copy
from typing import Union, List, Tuple

from tensortrade.features.feature_transformer import FeatureTransformer


[docs]class MinMaxNormalizer(FeatureTransformer): """A transformer for normalizing values within a feature pipeline by the column-wise extrema."""
[docs] def __init__(self, columns: Union[List[str], str, None] = None, feature_min=0, feature_max=1, inplace=True): """ Arguments: columns (optional): A list of column names to normalize. feature_min (optional): The minimum value in the range to scale to. feature_max (optional): The maximum value in the range to scale to. inplace (optional): If `False`, a new column will be added to the output for each input column. """ self._feature_min = feature_min self._feature_max = feature_max self._inplace = inplace self.columns = columns self._history = {}
[docs] def reset(self): self._history = {}
[docs] def transform_space(self, input_space: Space) -> Space: if self._inplace: return input_space output_space = copy(input_space) shape_x, *shape_y = input_space.shape columns = self.columns or range(len(shape_x)) output_space.shape = (shape_x + len(columns), *shape_y) for _ in columns: output_space.low = np.append(output_space.low, self._feature_min) output_space.high = np.append(output_space.high, self._feature_max) return output_space
[docs] def transform(self, X: pd.DataFrame) -> pd.DataFrame: if self.columns is None: self.columns = list(X.columns) for column in self.columns: prev_extrema = self._history.get(column, {'min': np.inf, 'max': -np.inf}) curr_min = min(X[column].min(), prev_extrema['min']) curr_max = max(X[column].max(), prev_extrema['max']) self._history[column] = {'min': curr_min, 'max': curr_max} scale = (self._feature_max - self._feature_min) + self._feature_min normalized_column = (X[column] - curr_min) / (curr_max - curr_min + 1E-9) * scale if self._inplace: X[column] = normalized_column else: X[f'{column}_minmax_{self._feature_min}_{self._feature_max}'] = normalized_column return X