KiTE

KiTE contains utilities to validate and calidrate supervised machine learning models.

Main Features

Here are the major utilities provided by the package:

  • Metrics to test if local bias is statistically significant within the given model
  • Calibration utilities to reduce local bias
  • Diffusion Map utilities to transform euclidean distance metrics into a diffusion space

Example Notebooks

We created Example Notebooks to showcase basic examples and applications of this library.

 1__docformat__ = "restructuredtext"
 2__doc__ = """
 3KiTE contains utilities to validate and calidrate supervised machine learning models.
 4
 5Main Features
 6-------------
 7Here are the major utilities provided by the package:
 8- Metrics to test if local bias is statistically significant within the given model
 9- Calibration utilities to reduce local bias
10- Diffusion Map utilities to transform euclidean distance metrics into a diffusion space
11
12
13Example Notebooks
14-----------------
15We created [Example Notebooks](https://github.com/A-Good-System-for-Smart-Cities/KiTE-utils/tree/main/notebooks) to showcase basic examples and applications of this library.
16
17"""
18
19import decorator
20import numpy as np
21
22none_arg_msg = "A given argument was None."
23
24
25@decorator.decorator
26def no_none_arg(f, *args, **kwargs):
27    is_none = [_ is None for _ in args if not isinstance(_, np.ndarray)]
28    if len(is_none) == 0 or True in is_none:
29        raise ValueError(none_arg_msg)
30    return f(*args, **kwargs)