import numpy as np
import pandas as pd
from desdeo_problem.surrogatemodels.SurrogateModels import BaseRegressor, ModelError
[docs]class LipschitzianRegressor(BaseRegressor):
def __init__(self, L: float = None):
self.L: float = L
self.X: np.ndarray = None
self.y: np.ndarray = None
[docs] def fit(self, X, y):
if isinstance(X, (pd.DataFrame, pd.Series)):
X = X.values
if isinstance(y, (pd.DataFrame, pd.Series)):
y = y.values.reshape(-1, 1)
# Make a 2-D array if needed
y = np.atleast_1d(y)
if y.ndim == 1:
y = y.reshape(-1, 1)
self_dist_x = self.self_distance(X)
self_dist_y = self.self_distance(y)
with np.errstate(divide="ignore", invalid="ignore"):
delta = np.true_divide(self_dist_y, self_dist_x)
delta[~np.isfinite(delta)] = -np.inf
L = delta.max()
self.X = X
self.y = y
self.L = L
[docs] def predict(self, X):
dist = self.distance(X, self.X)
y_low = (self.y - self.L * dist).max(axis=0)
y_high = (self.y + self.L * dist).min(axis=0)
y_mean = (y_low + y_high) / 2
y_delta = np.abs((y_high - y_low) / 2)
return (y_mean, y_delta)
[docs] def self_distance(self, arr):
if arr.ndim == 1:
dist = np.abs(np.subtract(arr[None, :], arr[:, None]))
elif arr.ndim == 2:
dist = np.sum(np.abs(np.subtract(arr[None, :, :], arr[:, None, :])), axis=2)
else:
msg = (
f"Array of wrong dimension. Expected dimension = 1 or 2. Recieved "
f"dimension = {arr.ndim}"
)
raise ModelError(msg)
return dist
[docs] def distance(self, array1, array2):
if array1.ndim == 1:
array1 = array1.reshape(-1, 1)
if array2.ndim == 1:
array2 = array2.reshape(-1, 1)
dist = np.sum(
np.abs(np.subtract(array1[None, :, :], array2[:, None, :])), axis=2
)
return dist
