File size: 23,252 Bytes
10ea30d
 
 
 
 
 
 
d426f8f
10ea30d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d426f8f
10ea30d
d426f8f
10ea30d
 
 
 
 
d426f8f
10ea30d
 
d426f8f
10ea30d
 
 
 
 
 
 
 
 
 
 
6ea45f3
10ea30d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bcef795
10ea30d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d426f8f
 
 
 
 
 
 
10ea30d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
---
license: mit
library_name: sklearn
tags:
- sklearn
- skops
- tabular-classification
model_file: churn.pkl
widget:
  structuredData:
    Contract:
    - Two year
    - Month-to-month
    - One year
    Dependents:
    - 'Yes'
    - 'No'
    - 'No'
    DeviceProtection:
    - 'No'
    - 'No'
    - 'Yes'
    InternetService:
    - Fiber optic
    - Fiber optic
    - DSL
    MonthlyCharges:
    - 79.05
    - 84.95
    - 68.8
    MultipleLines:
    - 'Yes'
    - 'Yes'
    - 'Yes'
    OnlineBackup:
    - 'No'
    - 'No'
    - 'Yes'
    OnlineSecurity:
    - 'Yes'
    - 'No'
    - 'Yes'
    PaperlessBilling:
    - 'No'
    - 'Yes'
    - 'No'
    Partner:
    - 'Yes'
    - 'Yes'
    - 'No'
    PaymentMethod:
    - Bank transfer (automatic)
    - Electronic check
    - Bank transfer (automatic)
    PhoneService:
    - 'Yes'
    - 'Yes'
    - 'Yes'
    SeniorCitizen:
    - 0
    - 0
    - 0
    StreamingMovies:
    - 'No'
    - 'No'
    - 'No'
    StreamingTV:
    - 'No'
    - 'Yes'
    - 'No'
    TechSupport:
    - 'No'
    - 'No'
    - 'Yes'
    TotalCharges:
    - 5730.7
    - 1378.25
    - 4111.35
    gender:
    - Female
    - Female
    - Male
    tenure:
    - 72
    - 16
    - 63
---

# Model description

This is a Logistic Regression model trained on churn dataset.

## Intended uses & limitations

This model is not ready to be used in production.

## Training Procedure

### Hyperparameters

The model is trained with below hyperparameters.

<details>
<summary> Click to expand </summary>

| Hyperparameter                             | Value                                                                             |
|--------------------------------------------|-----------------------------------------------------------------------------------|
| memory                                     |                                                                                   |
| steps                                      | [('preprocessor', ColumnTransformer(transformers=[('num',<br />                                 Pipeline(steps=[('imputer',<br />                                                  SimpleImputer(strategy='median')),<br />                                                 ('std_scaler',<br />                                                  StandardScaler())]),<br />                                 ['MonthlyCharges', 'TotalCharges', 'tenure']),<br />                                ('cat', OneHotEncoder(handle_unknown='ignore'),<br />                                 ['SeniorCitizen', 'gender', 'Partner',<br />                                  'Dependents', 'PhoneService', 'MultipleLines',<br />                                  'InternetService', 'OnlineSecurity',<br />                                  'OnlineBackup', 'DeviceProtection',<br />                                  'TechSupport', 'StreamingTV',<br />                                  'StreamingMovies', 'Contract',<br />                                  'PaperlessBilling', 'PaymentMethod'])])), ('classifier', LogisticRegression(class_weight='balanced', max_iter=300))]                                                                                   |
| verbose                                    | False                                                                             |
| preprocessor                               | ColumnTransformer(transformers=[('num',<br />                                 Pipeline(steps=[('imputer',<br />                                                  SimpleImputer(strategy='median')),<br />                                                 ('std_scaler',<br />                                                  StandardScaler())]),<br />                                 ['MonthlyCharges', 'TotalCharges', 'tenure']),<br />                                ('cat', OneHotEncoder(handle_unknown='ignore'),<br />                                 ['SeniorCitizen', 'gender', 'Partner',<br />                                  'Dependents', 'PhoneService', 'MultipleLines',<br />                                  'InternetService', 'OnlineSecurity',<br />                                  'OnlineBackup', 'DeviceProtection',<br />                                  'TechSupport', 'StreamingTV',<br />                                  'StreamingMovies', 'Contract',<br />                                  'PaperlessBilling', 'PaymentMethod'])])                                                                                   |
| classifier                                 | LogisticRegression(class_weight='balanced', max_iter=300)                         |
| preprocessor__n_jobs                       |                                                                                   |
| preprocessor__remainder                    | drop                                                                              |
| preprocessor__sparse_threshold             | 0.3                                                                               |
| preprocessor__transformer_weights          |                                                                                   |
| preprocessor__transformers                 | [('num', Pipeline(steps=[('imputer', SimpleImputer(strategy='median')),<br />                ('std_scaler', StandardScaler())]), ['MonthlyCharges', 'TotalCharges', 'tenure']), ('cat', OneHotEncoder(handle_unknown='ignore'), ['SeniorCitizen', 'gender', 'Partner', 'Dependents', 'PhoneService', 'MultipleLines', 'InternetService', 'OnlineSecurity', 'OnlineBackup', 'DeviceProtection', 'TechSupport', 'StreamingTV', 'StreamingMovies', 'Contract', 'PaperlessBilling', 'PaymentMethod'])]                                                                                   |
| preprocessor__verbose                      | False                                                                             |
| preprocessor__verbose_feature_names_out    | True                                                                              |
| preprocessor__num                          | Pipeline(steps=[('imputer', SimpleImputer(strategy='median')),<br />                ('std_scaler', StandardScaler())])                                                                                   |
| preprocessor__cat                          | OneHotEncoder(handle_unknown='ignore')                                            |
| preprocessor__num__memory                  |                                                                                   |
| preprocessor__num__steps                   | [('imputer', SimpleImputer(strategy='median')), ('std_scaler', StandardScaler())] |
| preprocessor__num__verbose                 | False                                                                             |
| preprocessor__num__imputer                 | SimpleImputer(strategy='median')                                                  |
| preprocessor__num__std_scaler              | StandardScaler()                                                                  |
| preprocessor__num__imputer__add_indicator  | False                                                                             |
| preprocessor__num__imputer__copy           | True                                                                              |
| preprocessor__num__imputer__fill_value     |                                                                                   |
| preprocessor__num__imputer__missing_values | nan                                                                               |
| preprocessor__num__imputer__strategy       | median                                                                            |
| preprocessor__num__imputer__verbose        | 0                                                                                 |
| preprocessor__num__std_scaler__copy        | True                                                                              |
| preprocessor__num__std_scaler__with_mean   | True                                                                              |
| preprocessor__num__std_scaler__with_std    | True                                                                              |
| preprocessor__cat__categories              | auto                                                                              |
| preprocessor__cat__drop                    |                                                                                   |
| preprocessor__cat__dtype                   | <class 'numpy.float64'>                                                           |
| preprocessor__cat__handle_unknown          | ignore                                                                            |
| preprocessor__cat__sparse                  | True                                                                              |
| classifier__C                              | 1.0                                                                               |
| classifier__class_weight                   | balanced                                                                          |
| classifier__dual                           | False                                                                             |
| classifier__fit_intercept                  | True                                                                              |
| classifier__intercept_scaling              | 1                                                                                 |
| classifier__l1_ratio                       |                                                                                   |
| classifier__max_iter                       | 300                                                                               |
| classifier__multi_class                    | auto                                                                              |
| classifier__n_jobs                         |                                                                                   |
| classifier__penalty                        | l2                                                                                |
| classifier__random_state                   |                                                                                   |
| classifier__solver                         | lbfgs                                                                             |
| classifier__tol                            | 0.0001                                                                            |
| classifier__verbose                        | 0                                                                                 |
| classifier__warm_start                     | False                                                                             |

</details>

### Model Plot

The model plot is below.

<style>#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 {color: black;background-color: white;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 pre{padding: 0;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-toggleable {background-color: white;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 label.sk-toggleable__label-arrow:before {content: "▸";float: left;margin-right: 0.25em;color: #696969;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: "▾";}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-estimator:hover {background-color: #d4ebff;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel-item::after {content: "";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-serial::before {content: "";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-item {z-index: 1;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel::before {content: "";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 2em;bottom: 0;left: 50%;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel-item {display: flex;flex-direction: column;position: relative;background-color: white;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-parallel-item:only-child::after {width: 0;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;position: relative;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-label label {font-family: monospace;font-weight: bold;background-color: white;display: inline-block;line-height: 1.2em;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-label-container {position: relative;z-index: 2;text-align: center;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-f0122ce0-64cb-41b3-8d66-0b116516efc3 div.sk-text-repr-fallback {display: none;}</style><div id="sk-f0122ce0-64cb-41b3-8d66-0b116516efc3" class="sk-top-container" style="overflow: auto;"><div class="sk-text-repr-fallback"><pre>Pipeline(steps=[(&#x27;preprocessor&#x27;,ColumnTransformer(transformers=[(&#x27;num&#x27;,Pipeline(steps=[(&#x27;imputer&#x27;,SimpleImputer(strategy=&#x27;median&#x27;)),(&#x27;std_scaler&#x27;,StandardScaler())]),[&#x27;MonthlyCharges&#x27;,&#x27;TotalCharges&#x27;, &#x27;tenure&#x27;]),(&#x27;cat&#x27;,OneHotEncoder(handle_unknown=&#x27;ignore&#x27;),[&#x27;SeniorCitizen&#x27;, &#x27;gender&#x27;,&#x27;Partner&#x27;, &#x27;Dependents&#x27;,&#x27;PhoneService&#x27;,&#x27;MultipleLines&#x27;,&#x27;InternetService&#x27;,&#x27;OnlineSecurity&#x27;,&#x27;OnlineBackup&#x27;,&#x27;DeviceProtection&#x27;,&#x27;TechSupport&#x27;, &#x27;StreamingTV&#x27;,&#x27;StreamingMovies&#x27;,&#x27;Contract&#x27;,&#x27;PaperlessBilling&#x27;,&#x27;PaymentMethod&#x27;])])),(&#x27;classifier&#x27;,LogisticRegression(class_weight=&#x27;balanced&#x27;, max_iter=300))])</pre><b>Please rerun this cell to show the HTML repr or trust the notebook.</b></div><div class="sk-container" hidden><div class="sk-item sk-dashed-wrapped"><div class="sk-label-container"><div class="sk-label sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="193bb424-11e4-4240-a49c-2b9ff9c16021" type="checkbox" ><label for="193bb424-11e4-4240-a49c-2b9ff9c16021" class="sk-toggleable__label sk-toggleable__label-arrow">Pipeline</label><div class="sk-toggleable__content"><pre>Pipeline(steps=[(&#x27;preprocessor&#x27;,ColumnTransformer(transformers=[(&#x27;num&#x27;,Pipeline(steps=[(&#x27;imputer&#x27;,SimpleImputer(strategy=&#x27;median&#x27;)),(&#x27;std_scaler&#x27;,StandardScaler())]),[&#x27;MonthlyCharges&#x27;,&#x27;TotalCharges&#x27;, &#x27;tenure&#x27;]),(&#x27;cat&#x27;,OneHotEncoder(handle_unknown=&#x27;ignore&#x27;),[&#x27;SeniorCitizen&#x27;, &#x27;gender&#x27;,&#x27;Partner&#x27;, &#x27;Dependents&#x27;,&#x27;PhoneService&#x27;,&#x27;MultipleLines&#x27;,&#x27;InternetService&#x27;,&#x27;OnlineSecurity&#x27;,&#x27;OnlineBackup&#x27;,&#x27;DeviceProtection&#x27;,&#x27;TechSupport&#x27;, &#x27;StreamingTV&#x27;,&#x27;StreamingMovies&#x27;,&#x27;Contract&#x27;,&#x27;PaperlessBilling&#x27;,&#x27;PaymentMethod&#x27;])])),(&#x27;classifier&#x27;,LogisticRegression(class_weight=&#x27;balanced&#x27;, max_iter=300))])</pre></div></div></div><div class="sk-serial"><div class="sk-item sk-dashed-wrapped"><div class="sk-label-container"><div class="sk-label sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="54004708-11cd-4f85-bff3-744af144ae72" type="checkbox" ><label for="54004708-11cd-4f85-bff3-744af144ae72" class="sk-toggleable__label sk-toggleable__label-arrow">preprocessor: ColumnTransformer</label><div class="sk-toggleable__content"><pre>ColumnTransformer(transformers=[(&#x27;num&#x27;,Pipeline(steps=[(&#x27;imputer&#x27;,SimpleImputer(strategy=&#x27;median&#x27;)),(&#x27;std_scaler&#x27;,StandardScaler())]),[&#x27;MonthlyCharges&#x27;, &#x27;TotalCharges&#x27;, &#x27;tenure&#x27;]),(&#x27;cat&#x27;, OneHotEncoder(handle_unknown=&#x27;ignore&#x27;),[&#x27;SeniorCitizen&#x27;, &#x27;gender&#x27;, &#x27;Partner&#x27;,&#x27;Dependents&#x27;, &#x27;PhoneService&#x27;, &#x27;MultipleLines&#x27;,&#x27;InternetService&#x27;, &#x27;OnlineSecurity&#x27;,&#x27;OnlineBackup&#x27;, &#x27;DeviceProtection&#x27;,&#x27;TechSupport&#x27;, &#x27;StreamingTV&#x27;,&#x27;StreamingMovies&#x27;, &#x27;Contract&#x27;,&#x27;PaperlessBilling&#x27;, &#x27;PaymentMethod&#x27;])])</pre></div></div></div><div class="sk-parallel"><div class="sk-parallel-item"><div class="sk-item"><div class="sk-label-container"><div class="sk-label sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="53cbe948-0bd7-4512-874e-7c0e8287ebf2" type="checkbox" ><label for="53cbe948-0bd7-4512-874e-7c0e8287ebf2" class="sk-toggleable__label sk-toggleable__label-arrow">num</label><div class="sk-toggleable__content"><pre>[&#x27;MonthlyCharges&#x27;, &#x27;TotalCharges&#x27;, &#x27;tenure&#x27;]</pre></div></div></div><div class="sk-serial"><div class="sk-item"><div class="sk-serial"><div class="sk-item"><div class="sk-estimator sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="9748155a-6575-4ba1-b5a2-9171c6ac1a11" type="checkbox" ><label for="9748155a-6575-4ba1-b5a2-9171c6ac1a11" class="sk-toggleable__label sk-toggleable__label-arrow">SimpleImputer</label><div class="sk-toggleable__content"><pre>SimpleImputer(strategy=&#x27;median&#x27;)</pre></div></div></div><div class="sk-item"><div class="sk-estimator sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="27303a89-9235-4743-862c-fa1959656bb7" type="checkbox" ><label for="27303a89-9235-4743-862c-fa1959656bb7" class="sk-toggleable__label sk-toggleable__label-arrow">StandardScaler</label><div class="sk-toggleable__content"><pre>StandardScaler()</pre></div></div></div></div></div></div></div></div><div class="sk-parallel-item"><div class="sk-item"><div class="sk-label-container"><div class="sk-label sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="0a07f5b9-db03-4bf5-bc2c-9b3f60e6ab16" type="checkbox" ><label for="0a07f5b9-db03-4bf5-bc2c-9b3f60e6ab16" class="sk-toggleable__label sk-toggleable__label-arrow">cat</label><div class="sk-toggleable__content"><pre>[&#x27;SeniorCitizen&#x27;, &#x27;gender&#x27;, &#x27;Partner&#x27;, &#x27;Dependents&#x27;, &#x27;PhoneService&#x27;, &#x27;MultipleLines&#x27;, &#x27;InternetService&#x27;, &#x27;OnlineSecurity&#x27;, &#x27;OnlineBackup&#x27;, &#x27;DeviceProtection&#x27;, &#x27;TechSupport&#x27;, &#x27;StreamingTV&#x27;, &#x27;StreamingMovies&#x27;, &#x27;Contract&#x27;, &#x27;PaperlessBilling&#x27;, &#x27;PaymentMethod&#x27;]</pre></div></div></div><div class="sk-serial"><div class="sk-item"><div class="sk-estimator sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="d985852a-65b0-4b77-897a-82c0ef3fa365" type="checkbox" ><label for="d985852a-65b0-4b77-897a-82c0ef3fa365" class="sk-toggleable__label sk-toggleable__label-arrow">OneHotEncoder</label><div class="sk-toggleable__content"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div></div></div></div></div></div></div></div><div class="sk-item"><div class="sk-estimator sk-toggleable"><input class="sk-toggleable__control sk-hidden--visually" id="050e23d8-6e98-4cfa-9ff8-cc01091c6a1f" type="checkbox" ><label for="050e23d8-6e98-4cfa-9ff8-cc01091c6a1f" class="sk-toggleable__label sk-toggleable__label-arrow">LogisticRegression</label><div class="sk-toggleable__content"><pre>LogisticRegression(class_weight=&#x27;balanced&#x27;, max_iter=300)</pre></div></div></div></div></div></div></div>

## Evaluation Results

You can find the details about evaluation process and the evaluation results.



| Metric   |    Value |
|----------|----------|
| accuracy | 0.730305 |
| f1 score | 0.730305 |

# How to Get Started with the Model

Use the code below to get started with the model.

```python
import joblib
import json
import pandas as pd
clf = joblib.load(churn.pkl)
with open("config.json") as f:
    config = json.load(f)
clf.predict(pd.DataFrame.from_dict(config["sklearn"]["example_input"]))
```


# Model Card Authors

This model card is written by following authors:

skops_user

# Model Card Contact

You can contact the model card authors through following channels:
[More Information Needed]

# Citation

Below you can find information related to citation.

**BibTeX:**
```
bibtex
@inproceedings{...,year={2020}}
```


# Additional Content

## confusion_matrix

![confusion_matrix](confusion_matrix.png)