├── LICENSE
├── .gitignore
├── README.md
├── requirements_ale.txt
├── requirements.txt
├── src
└── linear_models
│ └── challenge_notebook.ipynb
└── data
├── PDP_ICE.csv
├── insurance.csv
└── performance.csv
/LICENSE:
--------------------------------------------------------------------------------
1 | MIT License
2 |
3 | Copyright (c) 2023 a-data-odyssey
4 |
5 | Permission is hereby granted, free of charge, to any person obtaining a copy
6 | of this software and associated documentation files (the "Software"), to deal
7 | in the Software without restriction, including without limitation the rights
8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9 | copies of the Software, and to permit persons to whom the Software is
10 | furnished to do so, subject to the following conditions:
11 |
12 | The above copyright notice and this permission notice shall be included in all
13 | copies or substantial portions of the Software.
14 |
15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21 | SOFTWARE.
22 |
--------------------------------------------------------------------------------
/.gitignore:
--------------------------------------------------------------------------------
1 | # Byte-compiled / optimized / DLL files
2 | __pycache__/
3 | *.py[cod]
4 | *$py.class
5 |
6 | # C extensions
7 | *.so
8 |
9 | # Distribution / packaging
10 | .Python
11 | build/
12 | develop-eggs/
13 | dist/
14 | downloads/
15 | eggs/
16 | .eggs/
17 | lib/
18 | lib64/
19 | parts/
20 | sdist/
21 | var/
22 | wheels/
23 | share/python-wheels/
24 | *.egg-info/
25 | .installed.cfg
26 | *.egg
27 | MANIFEST
28 |
29 | # PyInstaller
30 | # Usually these files are written by a python script from a template
31 | # before PyInstaller builds the exe, so as to inject date/other infos into it.
32 | *.manifest
33 | *.spec
34 |
35 | # Installer logs
36 | pip-log.txt
37 | pip-delete-this-directory.txt
38 |
39 | # Unit test / coverage reports
40 | htmlcov/
41 | .tox/
42 | .nox/
43 | .coverage
44 | .coverage.*
45 | .cache
46 | nosetests.xml
47 | coverage.xml
48 | *.cover
49 | *.py,cover
50 | .hypothesis/
51 | .pytest_cache/
52 | cover/
53 |
54 | # Translations
55 | *.mo
56 | *.pot
57 |
58 | # Django stuff:
59 | *.log
60 | local_settings.py
61 | db.sqlite3
62 | db.sqlite3-journal
63 |
64 | # Flask stuff:
65 | instance/
66 | .webassets-cache
67 |
68 | # Scrapy stuff:
69 | .scrapy
70 |
71 | # Sphinx documentation
72 | docs/_build/
73 |
74 | # PyBuilder
75 | .pybuilder/
76 | target/
77 |
78 | # Jupyter Notebook
79 | .ipynb_checkpoints
80 |
81 | # IPython
82 | profile_default/
83 | ipython_config.py
84 |
85 | # pyenv
86 | # For a library or package, you might want to ignore these files since the code is
87 | # intended to run in multiple environments; otherwise, check them in:
88 | # .python-version
89 |
90 | # pipenv
91 | # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
92 | # However, in case of collaboration, if having platform-specific dependencies or dependencies
93 | # having no cross-platform support, pipenv may install dependencies that don't work, or not
94 | # install all needed dependencies.
95 | #Pipfile.lock
96 |
97 | # poetry
98 | # Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
99 | # This is especially recommended for binary packages to ensure reproducibility, and is more
100 | # commonly ignored for libraries.
101 | # https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
102 | #poetry.lock
103 |
104 | # pdm
105 | # Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
106 | #pdm.lock
107 | # pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
108 | # in version control.
109 | # https://pdm.fming.dev/#use-with-ide
110 | .pdm.toml
111 |
112 | # PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
113 | __pypackages__/
114 |
115 | # Celery stuff
116 | celerybeat-schedule
117 | celerybeat.pid
118 |
119 | # SageMath parsed files
120 | *.sage.py
121 |
122 | # Environments
123 | .env
124 | .venv
125 | env/
126 | venv/
127 | ENV/
128 | env.bak/
129 | venv.bak/
130 |
131 | # Spyder project settings
132 | .spyderproject
133 | .spyproject
134 |
135 | # Rope project settings
136 | .ropeproject
137 |
138 | # mkdocs documentation
139 | /site
140 |
141 | # mypy
142 | .mypy_cache/
143 | .dmypy.json
144 | dmypy.json
145 |
146 | # Pyre type checker
147 | .pyre/
148 |
149 | # pytype static type analyzer
150 | .pytype/
151 |
152 | # Cython debug symbols
153 | cython_debug/
154 |
155 | # PyCharm
156 | # JetBrains specific template is maintained in a separate JetBrains.gitignore that can
157 | # be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
158 | # and can be added to the global gitignore or merged into this file. For a more nuclear
159 | # option (not recommended) you can uncomment the following to ignore the entire idea folder.
160 | #.idea/
161 |
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # XAI-tutorial
2 | A course on explainable AI with Python. You can access the course [here](https://adataodyssey.com/courses/xai-with-python/).
3 |
4 | ### File Structure and Descriptions
5 | The resources in the `src` folder are as follows:
6 |
7 | | Folder | File | Dataset | Chapter | Description |
8 | | ------------------ | -------------------------------- | ----------------- | ------- | ----------------------------------------------------------------------------------------------------- |
9 | | intro | human_friendly_explanations.ipynb| insurance.csv | 4 | Using SHAP plots as the basis for human-friendly explanations |
10 | | linear_models | power_of_linear_models.ipynb | performance.csv | 5 | Demonstrating how feature engineering allows linear models to model non-linear relationships |
11 | | | feature_engineering.ipynb | credit_score.csv | 6 | Interpretable feature engineering methods |
12 | | | feature_clustering.ipynb | credit_score.csv | 8 | Feature selection with hierarchical feature clustering |
13 | | | linear_regression.ipynb | credit_score.csv | 9 | Explaining linear regression to a non-technical audience |
14 | | | challenge_notebook.ipynb | credit_score.csv | 5-9 | Logistic regression model for Part 2 challenge |
15 | | model_agnostic | feature_importance.ipynb | credit_score.csv | 11 | Calculating PFI scores from scratch |
16 | | | pdp_and_ice.ipynb | PDP_ICE.csv | 12 | Applying PDPs and ICE plots with scikit-learn |
17 | | | ale.ipynb | abalone.data | 13 | Applying ALEs with Alibi Explained |
18 | | | h_stat.ipynb | abalone.data | 14 | Applying Friedman's H-stat with artemis |
19 | | | lime.ipynb | abalone.data | 15 | Applying lime |
20 | | | shap.ipynb | credit_score.csv | 16 | Applying SHAP |
21 | | misc | transaction_data_generator.ipynb | - | - | Used to generate credit_score dataset |
22 |
23 | ### Datasets
24 |
25 | The datasets used by the files can be found in the `data` folder. More information about the datasets is available at these links:
26 |
27 | - [credit_score.csv](https://www.kaggle.com/datasets/conorsully1/credit-score)
28 | - [insurance.csv](https://www.kaggle.com/datasets/mirichoi0218/insurance)
29 | - [abalone.data](https://archive.ics.uci.edu/ml/datasets/Abalone)
30 | - [PDP_ICE.csv](https://www.kaggle.com/datasets/conorsully1/pdp-and-ice-plots)
31 | - [performance.csv](https://www.kaggle.com/datasets/conorsully1/annual-promotions)
32 |
33 | ### Packages and Versions
34 |
35 | For the majority of the course, we use:
36 |
37 | - Python 3.11.5
38 | - `requirements.txt`
39 |
40 | For Chapter 13's notebook, `ale.ipynb`, we use:
41 |
42 | - Python 3.9.12
43 | - `requirements_ale.txt`
44 |
--------------------------------------------------------------------------------
/requirements_ale.txt:
--------------------------------------------------------------------------------
1 | alibi==0.9.4
2 | annotated-types==0.6.0
3 | anyio==4.0.0
4 | appnope==0.1.3
5 | argon2-cffi==23.1.0
6 | argon2-cffi-bindings==21.2.0
7 | arrow==1.3.0
8 | asttokens==2.4.0
9 | async-lru==2.0.4
10 | attrs==23.1.0
11 | Babel==2.13.0
12 | backcall==0.2.0
13 | beautifulsoup4==4.12.2
14 | bleach==6.1.0
15 | blis==0.7.11
16 | catalogue==2.0.10
17 | catboost==1.2.2
18 | celluloid==0.2.0
19 | certifi==2023.7.22
20 | cffi==1.16.0
21 | charset-normalizer==3.3.0
22 | click==8.1.7
23 | cloudpathlib==0.16.0
24 | cloudpickle==2.2.1
25 | comm==0.1.4
26 | confection==0.1.3
27 | contourpy==1.1.1
28 | cycler==0.12.1
29 | cymem==2.0.8
30 | dalex==1.6.0
31 | debugpy==1.8.0
32 | decorator==5.1.1
33 | defusedxml==0.7.1
34 | dill==0.3.7
35 | exceptiongroup==1.1.3
36 | executing==2.0.0
37 | fastjsonschema==2.18.1
38 | filelock==3.13.1
39 | fonttools==4.43.1
40 | fqdn==1.5.1
41 | fsspec==2023.10.0
42 | graphviz==0.20.1
43 | huggingface-hub==0.19.3
44 | idna==3.4
45 | imageio==2.32.0
46 | importlib-metadata==6.8.0
47 | importlib-resources==6.1.0
48 | ipykernel==6.25.2
49 | ipython==8.16.1
50 | ipython-genutils==0.2.0
51 | ipywidgets==8.1.1
52 | isoduration==20.11.0
53 | jedi==0.19.1
54 | Jinja2==3.1.2
55 | joblib==1.3.2
56 | json5==0.9.14
57 | jsonpointer==2.4
58 | jsonschema==4.19.1
59 | jsonschema-specifications==2023.7.1
60 | jupyter==1.0.0
61 | jupyter-console==6.6.3
62 | jupyter-events==0.7.0
63 | jupyter-lsp==2.2.0
64 | jupyter_client==8.4.0
65 | jupyter_core==5.4.0
66 | jupyter_server==2.7.3
67 | jupyter_server_terminals==0.4.4
68 | jupyterlab==4.0.7
69 | jupyterlab-pygments==0.2.2
70 | jupyterlab-widgets==3.0.9
71 | jupyterlab_server==2.25.0
72 | kiwisolver==1.4.5
73 | langcodes==3.3.0
74 | lazy_loader==0.3
75 | llvmlite==0.41.0
76 | MarkupSafe==2.1.3
77 | matplotlib==3.8.0
78 | matplotlib-inline==0.1.6
79 | mistune==3.0.2
80 | murmurhash==1.0.10
81 | nbclient==0.8.0
82 | nbconvert==7.9.2
83 | nbformat==5.9.2
84 | nest-asyncio==1.5.8
85 | networkx==2.8.8
86 | notebook==7.0.5
87 | notebook_shim==0.2.3
88 | numba==0.58.0
89 | numpy==1.25.0
90 | overrides==7.4.0
91 | packaging==23.2
92 | pandas==1.5.3
93 | pandocfilters==1.5.0
94 | parso==0.8.3
95 | patsy==0.5.3
96 | pexpect==4.8.0
97 | pickleshare==0.7.5
98 | Pillow==9.5.0
99 | platformdirs==3.11.0
100 | plotly==5.18.0
101 | preshed==3.0.9
102 | prometheus-client==0.17.1
103 | prompt-toolkit==3.0.39
104 | psutil==5.9.5
105 | ptyprocess==0.7.0
106 | pure-eval==0.2.2
107 | pycparser==2.21
108 | pydantic==2.5.1
109 | pydantic_core==2.14.3
110 | Pygments==2.16.1
111 | pyparsing==3.1.1
112 | python-dateutil==2.8.2
113 | python-json-logger==2.0.7
114 | pytz==2023.3.post1
115 | PyWavelets==1.4.1
116 | PyYAML==6.0.1
117 | pyzmq==25.1.1
118 | qtconsole==5.4.4
119 | QtPy==2.4.0
120 | referencing==0.30.2
121 | regex==2023.10.3
122 | requests==2.31.0
123 | rfc3339-validator==0.1.4
124 | rfc3986-validator==0.1.1
125 | rpds-py==0.10.6
126 | safetensors==0.4.0
127 | scikit-image==0.21.0
128 | scikit-learn==1.3.1
129 | scipy==1.11.3
130 | seaborn==0.13.0
131 | Send2Trash==1.8.2
132 | shap==0.43.0
133 | six==1.16.0
134 | slicer==0.0.7
135 | smart-open==6.4.0
136 | sniffio==1.3.0
137 | soupsieve==2.5
138 | spacy==3.7.2
139 | spacy-legacy==3.0.12
140 | spacy-loggers==1.0.5
141 | spacy-lookups-data==1.0.5
142 | srsly==2.4.8
143 | stack-data==0.6.3
144 | statsmodels==0.14.0
145 | tenacity==8.2.3
146 | terminado==0.17.1
147 | thinc==8.2.1
148 | threadpoolctl==3.2.0
149 | tifffile==2023.9.26
150 | tinycss2==1.2.1
151 | tokenizers==0.15.0
152 | tomli==2.0.1
153 | tornado==6.3.3
154 | tqdm==4.66.1
155 | traitlets==5.11.2
156 | transformers==4.35.2
157 | typer==0.9.0
158 | types-python-dateutil==2.8.19.14
159 | typing_extensions==4.8.0
160 | tzdata==2023.3
161 | ucimlrepo==0.0.2
162 | uri-template==1.3.0
163 | urllib3==2.0.6
164 | wasabi==1.1.2
165 | wcwidth==0.2.8
166 | weasel==0.3.4
167 | webcolors==1.13
168 | webencodings==0.5.1
169 | websocket-client==1.6.4
170 | widgetsnbextension==4.0.9
171 | xgboost==2.0.1
172 | zipp==3.17.0
173 |
--------------------------------------------------------------------------------
/requirements.txt:
--------------------------------------------------------------------------------
1 | absl-py==2.0.0
2 | alibi==0.9.4
3 | annotated-types==0.6.0
4 | anyio==4.0.0
5 | appnope==0.1.3
6 | argon2-cffi==23.1.0
7 | argon2-cffi-bindings==21.2.0
8 | arrow==1.3.0
9 | asttokens==2.4.0
10 | astunparse==1.6.3
11 | async-lru==2.0.4
12 | attrs==23.1.0
13 | Babel==2.13.0
14 | backcall==0.2.0
15 | beautifulsoup4==4.12.2
16 | bleach==6.1.0
17 | blis==0.7.11
18 | cachetools==5.3.2
19 | catalogue==2.0.10
20 | catboost==1.2.2
21 | celluloid==0.2.0
22 | certifi==2023.7.22
23 | cffi==1.16.0
24 | charset-normalizer==3.3.0
25 | click==8.1.7
26 | cloudpathlib==0.16.0
27 | cloudpickle==2.2.1
28 | comm==0.1.4
29 | confection==0.1.3
30 | contourpy==1.1.1
31 | cycler==0.12.1
32 | cymem==2.0.8
33 | dalex==1.6.0
34 | debugpy==1.8.0
35 | decorator==5.1.1
36 | defusedxml==0.7.1
37 | dill==0.3.7
38 | dm-tree==0.1.8
39 | exceptiongroup==1.1.3
40 | executing==2.0.0
41 | fastjsonschema==2.18.1
42 | filelock==3.13.1
43 | flatbuffers==23.5.26
44 | fonttools==4.43.1
45 | fqdn==1.5.1
46 | fsspec==2023.10.0
47 | gast==0.5.4
48 | google-auth==2.25.2
49 | google-auth-oauthlib==1.2.0
50 | google-pasta==0.2.0
51 | graphviz==0.20.1
52 | grpcio==1.60.0
53 | h5py==3.10.0
54 | huggingface-hub==0.19.3
55 | idna==3.4
56 | imageio==2.32.0
57 | importlib-metadata==6.8.0
58 | importlib-resources==6.1.0
59 | ipykernel==6.25.2
60 | ipython==8.16.1
61 | ipython-genutils==0.2.0
62 | ipywidgets==8.1.1
63 | isoduration==20.11.0
64 | jedi==0.19.1
65 | Jinja2==3.1.2
66 | joblib==1.3.2
67 | json5==0.9.14
68 | jsonpointer==2.4
69 | jsonschema==4.19.1
70 | jsonschema-specifications==2023.7.1
71 | jupyter==1.0.0
72 | jupyter-console==6.6.3
73 | jupyter-events==0.7.0
74 | jupyter-lsp==2.2.0
75 | jupyter_client==8.4.0
76 | jupyter_core==5.4.0
77 | jupyter_server==2.7.3
78 | jupyter_server_terminals==0.4.4
79 | jupyterlab==4.0.7
80 | jupyterlab-pygments==0.2.2
81 | jupyterlab-widgets==3.0.9
82 | jupyterlab_server==2.25.0
83 | kiwisolver==1.4.5
84 | langcodes==3.3.0
85 | lazy_loader==0.3
86 | libclang==16.0.6
87 | lime==0.2.0.1
88 | llvmlite==0.41.0
89 | Markdown==3.5.1
90 | markdown-it-py==3.0.0
91 | MarkupSafe==2.1.3
92 | matplotlib==3.8.0
93 | matplotlib-inline==0.1.6
94 | mdurl==0.1.2
95 | mistune==3.0.2
96 | ml-dtypes==0.2.0
97 | murmurhash==1.0.10
98 | namex==0.0.7
99 | nbclient==0.8.0
100 | nbconvert==7.9.2
101 | nbformat==5.9.2
102 | nest-asyncio==1.5.8
103 | networkx==2.8.8
104 | notebook==7.0.5
105 | notebook_shim==0.2.3
106 | numba==0.58.0
107 | numpy==1.25.0
108 | oauthlib==3.2.2
109 | opt-einsum==3.3.0
110 | overrides==7.4.0
111 | packaging==23.2
112 | pandas==1.5.3
113 | pandocfilters==1.5.0
114 | parso==0.8.3
115 | patsy==0.5.3
116 | pexpect==4.8.0
117 | pickleshare==0.7.5
118 | Pillow==9.5.0
119 | platformdirs==3.11.0
120 | plotly==5.18.0
121 | preshed==3.0.9
122 | prometheus-client==0.17.1
123 | prompt-toolkit==3.0.39
124 | protobuf==4.23.4
125 | psutil==5.9.5
126 | ptyprocess==0.7.0
127 | pure-eval==0.2.2
128 | pyartemis==0.1.5
129 | pyasn1==0.5.1
130 | pyasn1-modules==0.3.0
131 | pycparser==2.21
132 | pydantic==2.5.1
133 | pydantic_core==2.14.3
134 | Pygments==2.16.1
135 | pyparsing==3.1.1
136 | python-dateutil==2.8.2
137 | python-json-logger==2.0.7
138 | pytz==2023.3.post1
139 | PyWavelets==1.4.1
140 | PyYAML==6.0.1
141 | pyzmq==25.1.1
142 | qtconsole==5.4.4
143 | QtPy==2.4.0
144 | referencing==0.30.2
145 | regex==2023.10.3
146 | requests==2.31.0
147 | requests-oauthlib==1.3.1
148 | rfc3339-validator==0.1.4
149 | rfc3986-validator==0.1.1
150 | rich==13.7.0
151 | rpds-py==0.10.6
152 | rsa==4.9
153 | safetensors==0.4.0
154 | scikit-image==0.21.0
155 | scikit-learn==1.3.1
156 | scipy==1.11.3
157 | seaborn==0.13.0
158 | Send2Trash==1.8.2
159 | shap==0.43.0
160 | six==1.16.0
161 | slicer==0.0.7
162 | smart-open==6.4.0
163 | sniffio==1.3.0
164 | soupsieve==2.5
165 | spacy==3.7.2
166 | spacy-legacy==3.0.12
167 | spacy-loggers==1.0.5
168 | spacy-lookups-data==1.0.5
169 | srsly==2.4.8
170 | stack-data==0.6.3
171 | statsmodels==0.14.0
172 | tenacity==8.2.3
173 | tensorboard==2.15.1
174 | tensorboard-data-server==0.7.2
175 | tensorflow-estimator==2.15.0
176 | tensorflow-io-gcs-filesystem==0.34.0
177 | tensorflow-macos==2.15.0
178 | termcolor==2.4.0
179 | terminado==0.17.1
180 | thinc==8.2.1
181 | threadpoolctl==3.2.0
182 | tifffile==2023.9.26
183 | tinycss2==1.2.1
184 | tokenizers==0.15.0
185 | tomli==2.0.1
186 | tornado==6.3.3
187 | tqdm==4.66.1
188 | traitlets==5.11.2
189 | transformers==4.35.2
190 | typer==0.9.0
191 | types-python-dateutil==2.8.19.14
192 | typing_extensions==4.8.0
193 | tzdata==2023.3
194 | ucimlrepo==0.0.2
195 | uri-template==1.3.0
196 | urllib3==2.0.6
197 | wasabi==1.1.2
198 | wcwidth==0.2.8
199 | weasel==0.3.4
200 | webcolors==1.13
201 | webencodings==0.5.1
202 | websocket-client==1.6.4
203 | Werkzeug==3.0.1
204 | widgetsnbextension==4.0.9
205 | wrapt==1.14.1
206 | xgboost==2.0.1
207 | zipp==3.17.0
208 |
--------------------------------------------------------------------------------
/src/linear_models/challenge_notebook.ipynb:
--------------------------------------------------------------------------------
1 | {
2 | "cells": [
3 | {
4 | "cell_type": "code",
5 | "execution_count": 1,
6 | "metadata": {},
7 | "outputs": [],
8 | "source": [
9 | "import statsmodels.api as sm\n",
10 | "\n",
11 | "import matplotlib.pyplot as plt\n",
12 | "import seaborn as sns\n",
13 | "\n",
14 | "import numpy as np\n",
15 | "import pandas as pd\n",
16 | "\n",
17 | "from sklearn.preprocessing import MinMaxScaler\n",
18 | "from sklearn.preprocessing import StandardScaler\n",
19 | "import shap"
20 | ]
21 | },
22 | {
23 | "cell_type": "code",
24 | "execution_count": 2,
25 | "metadata": {},
26 | "outputs": [
27 | {
28 | "name": "stdout",
29 | "output_type": "stream",
30 | "text": [
31 | "152 304 304\n"
32 | ]
33 | }
34 | ],
35 | "source": [
36 | "# fetch dataset \n",
37 | "credit_score = pd.read_csv(\"../../data/credit_score.csv\")\n",
38 | "credit_score['R_EXPENDITURE_SQR'] = credit_score['R_EXPENDITURE']**2\n",
39 | "\n",
40 | "# data (as pandas dataframes) \n",
41 | "X = credit_score[['INCOME','R_DEBT_INCOME','R_EXPENDITURE','R_EXPENDITURE_SQR','R_EXPENDITURE_SAVINGS','R_ENTERTAINMENT','CAT_GAMBLING']].copy()\n",
42 | "\n",
43 | "# one-hot encoding\n",
44 | "X['GAMBLING_LOW'] = X['CAT_GAMBLING'].apply(lambda x: 1 if x == 'Low' else 0)\n",
45 | "X['GAMBLING_HIGH'] = X['CAT_GAMBLING'].apply(lambda x: 1 if x == 'High' else 0)\n",
46 | "X.drop(columns=['CAT_GAMBLING'], inplace=True)\n",
47 | "\n",
48 | "y = credit_score['DEFAULT']\n",
49 | "\n",
50 | "# get a balanced dataset\n",
51 | "n = np.sum(y)\n",
52 | "X = pd.concat([X[y==1],X[y==0].sample(n=n, random_state=0)])\n",
53 | "y = pd.concat([y[y==1],y[y==0].sample(n=n, random_state=0)])\n",
54 | "\n",
55 | "print(n,len(X),len(y))\n"
56 | ]
57 | },
58 | {
59 | "cell_type": "code",
60 | "execution_count": 3,
61 | "metadata": {},
62 | "outputs": [
63 | {
64 | "name": "stdout",
65 | "output_type": "stream",
66 | "text": [
67 | "Optimization terminated successfully.\n",
68 | " Current function value: 0.570557\n",
69 | " Iterations 6\n"
70 | ]
71 | },
72 | {
73 | "data": {
74 | "text/html": [
75 | "\n",
76 | "Logit Regression Results\n",
77 | "\n",
78 | " | Dep. Variable: | DEFAULT | No. Observations: | 304 | \n",
79 | "
\n",
80 | "\n",
81 | " | Model: | Logit | Df Residuals: | 296 | \n",
82 | "
\n",
83 | "\n",
84 | " | Method: | MLE | Df Model: | 7 | \n",
85 | "
\n",
86 | "\n",
87 | " | Date: | Tue, 24 Oct 2023 | Pseudo R-squ.: | 0.1769 | \n",
88 | "
\n",
89 | "\n",
90 | " | Time: | 12:36:23 | Log-Likelihood: | -173.45 | \n",
91 | "
\n",
92 | "\n",
93 | " | converged: | True | LL-Null: | -210.72 | \n",
94 | "
\n",
95 | "\n",
96 | " | Covariance Type: | nonrobust | LLR p-value: | 1.782e-13 | \n",
97 | "
\n",
98 | "
\n",
99 | "\n",
100 | "\n",
101 | " | coef | std err | z | P>|z| | [0.025 | 0.975] | \n",
102 | "
\n",
103 | "\n",
104 | " | INCOME | -3.14e-06 | 1.35e-06 | -2.326 | 0.020 | -5.79e-06 | -4.94e-07 | \n",
105 | "
\n",
106 | "\n",
107 | " | R_DEBT_INCOME | 0.1334 | 0.026 | 5.212 | 0.000 | 0.083 | 0.184 | \n",
108 | "
\n",
109 | "\n",
110 | " | R_EXPENDITURE | -14.1720 | 4.719 | -3.003 | 0.003 | -23.421 | -4.923 | \n",
111 | "
\n",
112 | "\n",
113 | " | R_EXPENDITURE_SQR | 18.8127 | 5.291 | 3.556 | 0.000 | 8.443 | 29.182 | \n",
114 | "
\n",
115 | "\n",
116 | " | R_EXPENDITURE_SAVINGS | 0.2304 | 0.075 | 3.070 | 0.002 | 0.083 | 0.377 | \n",
117 | "
\n",
118 | "\n",
119 | " | R_ENTERTAINMENT | 1.9536 | 1.963 | 0.995 | 0.320 | -1.894 | 5.801 | \n",
120 | "
\n",
121 | "\n",
122 | " | GAMBLING_LOW | 0.2502 | 0.442 | 0.566 | 0.572 | -0.617 | 1.117 | \n",
123 | "
\n",
124 | "\n",
125 | " | GAMBLING_HIGH | 0.5603 | 0.309 | 1.815 | 0.070 | -0.045 | 1.165 | \n",
126 | "
\n",
127 | "
"
128 | ],
129 | "text/latex": [
130 | "\\begin{center}\n",
131 | "\\begin{tabular}{lclc}\n",
132 | "\\toprule\n",
133 | "\\textbf{Dep. Variable:} & DEFAULT & \\textbf{ No. Observations: } & 304 \\\\\n",
134 | "\\textbf{Model:} & Logit & \\textbf{ Df Residuals: } & 296 \\\\\n",
135 | "\\textbf{Method:} & MLE & \\textbf{ Df Model: } & 7 \\\\\n",
136 | "\\textbf{Date:} & Tue, 24 Oct 2023 & \\textbf{ Pseudo R-squ.: } & 0.1769 \\\\\n",
137 | "\\textbf{Time:} & 12:36:23 & \\textbf{ Log-Likelihood: } & -173.45 \\\\\n",
138 | "\\textbf{converged:} & True & \\textbf{ LL-Null: } & -210.72 \\\\\n",
139 | "\\textbf{Covariance Type:} & nonrobust & \\textbf{ LLR p-value: } & 1.782e-13 \\\\\n",
140 | "\\bottomrule\n",
141 | "\\end{tabular}\n",
142 | "\\begin{tabular}{lcccccc}\n",
143 | " & \\textbf{coef} & \\textbf{std err} & \\textbf{z} & \\textbf{P$> |$z$|$} & \\textbf{[0.025} & \\textbf{0.975]} \\\\\n",
144 | "\\midrule\n",
145 | "\\textbf{INCOME} & -3.14e-06 & 1.35e-06 & -2.326 & 0.020 & -5.79e-06 & -4.94e-07 \\\\\n",
146 | "\\textbf{R\\_DEBT\\_INCOME} & 0.1334 & 0.026 & 5.212 & 0.000 & 0.083 & 0.184 \\\\\n",
147 | "\\textbf{R\\_EXPENDITURE} & -14.1720 & 4.719 & -3.003 & 0.003 & -23.421 & -4.923 \\\\\n",
148 | "\\textbf{R\\_EXPENDITURE\\_SQR} & 18.8127 & 5.291 & 3.556 & 0.000 & 8.443 & 29.182 \\\\\n",
149 | "\\textbf{R\\_EXPENDITURE\\_SAVINGS} & 0.2304 & 0.075 & 3.070 & 0.002 & 0.083 & 0.377 \\\\\n",
150 | "\\textbf{R\\_ENTERTAINMENT} & 1.9536 & 1.963 & 0.995 & 0.320 & -1.894 & 5.801 \\\\\n",
151 | "\\textbf{GAMBLING\\_LOW} & 0.2502 & 0.442 & 0.566 & 0.572 & -0.617 & 1.117 \\\\\n",
152 | "\\textbf{GAMBLING\\_HIGH} & 0.5603 & 0.309 & 1.815 & 0.070 & -0.045 & 1.165 \\\\\n",
153 | "\\bottomrule\n",
154 | "\\end{tabular}\n",
155 | "%\\caption{Logit Regression Results}\n",
156 | "\\end{center}"
157 | ],
158 | "text/plain": [
159 | "\n",
160 | "\"\"\"\n",
161 | " Logit Regression Results \n",
162 | "==============================================================================\n",
163 | "Dep. Variable: DEFAULT No. Observations: 304\n",
164 | "Model: Logit Df Residuals: 296\n",
165 | "Method: MLE Df Model: 7\n",
166 | "Date: Tue, 24 Oct 2023 Pseudo R-squ.: 0.1769\n",
167 | "Time: 12:36:23 Log-Likelihood: -173.45\n",
168 | "converged: True LL-Null: -210.72\n",
169 | "Covariance Type: nonrobust LLR p-value: 1.782e-13\n",
170 | "=========================================================================================\n",
171 | " coef std err z P>|z| [0.025 0.975]\n",
172 | "-----------------------------------------------------------------------------------------\n",
173 | "INCOME -3.14e-06 1.35e-06 -2.326 0.020 -5.79e-06 -4.94e-07\n",
174 | "R_DEBT_INCOME 0.1334 0.026 5.212 0.000 0.083 0.184\n",
175 | "R_EXPENDITURE -14.1720 4.719 -3.003 0.003 -23.421 -4.923\n",
176 | "R_EXPENDITURE_SQR 18.8127 5.291 3.556 0.000 8.443 29.182\n",
177 | "R_EXPENDITURE_SAVINGS 0.2304 0.075 3.070 0.002 0.083 0.377\n",
178 | "R_ENTERTAINMENT 1.9536 1.963 0.995 0.320 -1.894 5.801\n",
179 | "GAMBLING_LOW 0.2502 0.442 0.566 0.572 -0.617 1.117\n",
180 | "GAMBLING_HIGH 0.5603 0.309 1.815 0.070 -0.045 1.165\n",
181 | "=========================================================================================\n",
182 | "\"\"\""
183 | ]
184 | },
185 | "execution_count": 3,
186 | "metadata": {},
187 | "output_type": "execute_result"
188 | }
189 | ],
190 | "source": [
191 | "# fit model\n",
192 | "model = sm.Logit(y, X).fit()\n",
193 | "model.summary()"
194 | ]
195 | },
196 | {
197 | "cell_type": "code",
198 | "execution_count": 4,
199 | "metadata": {},
200 | "outputs": [
201 | {
202 | "name": "stdout",
203 | "output_type": "stream",
204 | "text": [
205 | "Model accuracy: 0.694078947368421\n"
206 | ]
207 | },
208 | {
209 | "data": {
210 | "text/html": [
211 | "\n",
212 | "\n",
225 | "
\n",
226 | " \n",
227 | " \n",
228 | " | Actual | \n",
229 | " 0 | \n",
230 | " 1 | \n",
231 | "
\n",
232 | " \n",
233 | " | Predicted | \n",
234 | " | \n",
235 | " | \n",
236 | "
\n",
237 | " \n",
238 | " \n",
239 | " \n",
240 | " | 0 | \n",
241 | " 112 | \n",
242 | " 53 | \n",
243 | "
\n",
244 | " \n",
245 | " | 1 | \n",
246 | " 40 | \n",
247 | " 99 | \n",
248 | "
\n",
249 | " \n",
250 | "
\n",
251 | "