import numpy as np
from typing import Dict, Callable, List
[docs]
class Discretizer(object):
[docs]
def __init__(self, data: np.ndarray, categorical_features: List[int], feature_names: List[str],
percentiles: List[int] = [25, 50, 75]) -> None:
"""
Initialize the discretizer.
Parameters
----------
data
Data to discretize
categorical_features
List of indices corresponding to the categorical columns. These features will not be discretized.
The other features will be considered continuous and therefore discretized.
feature_names
List with feature names
percentiles
Percentiles used for discretization
"""
self.to_discretize = ([x for x in range(data.shape[1]) if x not in categorical_features])
self.percentiles = percentiles
bins = self.bins(data)
bins = [np.unique(x) for x in bins]
self.names: Dict[int, list] = {}
self.lambdas: Dict[int, Callable] = {}
for feature, qts in zip(self.to_discretize, bins):
# get nb of borders (nb of bins - 1) and the feature name
n_bins = qts.shape[0]
name = feature_names[feature]
# create names for bins of discretized features
self.names[feature] = ['%s <= %.2f' % (name, qts[0])]
for i in range(n_bins - 1):
self.names[feature].append('%.2f < %s <= %.2f' % (qts[i], name, qts[i + 1]))
self.names[feature].append('%s > %.2f' % (name, qts[n_bins - 1]))
self.lambdas[feature] = lambda x, qts = qts: np.searchsorted(qts, x)
[docs]
def bins(self, data: np.ndarray) -> List[np.ndarray]:
"""
Parameters
----------
data
Data to discretize
Returns
-------
List with bin values for each feature that is discretized.
"""
bins = []
for feature in self.to_discretize:
qts = np.array(np.percentile(data[:, feature], self.percentiles))
bins.append(qts)
return bins
[docs]
def discretize(self, data: np.ndarray) -> np.ndarray:
"""
Parameters
----------
data
Data to discretize
Returns
-------
Discretized version of data with the same dimension.
"""
data_disc = data.copy()
for feature in self.lambdas:
if len(data.shape) == 1:
data_disc[feature] = int(self.lambdas[feature](data_disc[feature]))
else:
data_disc[:, feature] = self.lambdas[feature](data_disc[:, feature]).astype(int)
return data_disc