# check version
from pycaret.utils import version
version()

'2.3.1'

1. Loading Dataset 

from pycaret.datasets import get_data
data = get_data('insurance')
age sex bmi children smoker region charges
0 19 female 27.900 0 yes southwest 16884.92400
1 18 male 33.770 1 no southeast 1725.55230
2 28 male 33.000 3 no southeast 4449.46200
3 33 male 22.705 0 no northwest 21984.47061
4 32 male 28.880 0 no northwest 3866.85520 
import pandas as pd
url = 'https://raw.githubusercontent.com/davidrkearney/colab-notebooks/main/datasets/strokes_training.csv'
df = pd.read_csv(url, error_bad_lines=False)
df
id gender age hypertension heart_disease ever_married work_type Residence_type avg_glucose_level bmi smoking_status stroke
0 30669 Male 3.0 0 0 No children Rural 95.12 18.0 NaN 0
1 30468 Male 58.0 1 0 Yes Private Urban 87.96 39.2 never smoked 0
2 16523 Female 8.0 0 0 No Private Urban 110.89 17.6 NaN 0
3 56543 Female 70.0 0 0 Yes Private Rural 69.04 35.9 formerly smoked 0
4 46136 Male 14.0 0 0 No Never_worked Rural 161.28 19.1 NaN 0
... ... ... ... ... ... ... ... ... ... ... ... ...
43395 56196 Female 10.0 0 0 No children Urban 58.64 20.4 never smoked 0
43396 5450 Female 56.0 0 0 Yes Govt_job Urban 213.61 55.4 formerly smoked 0
43397 28375 Female 82.0 1 0 Yes Private Urban 91.94 28.9 formerly smoked 0
43398 27973 Male 40.0 0 0 Yes Private Urban 99.16 33.2 never smoked 0
43399 36271 Female 82.0 0 0 Yes Private Urban 79.48 20.6 never smoked 0

43400 rows × 12 columns

2. Initialize Setup 

from pycaret.regression import *
reg1 = setup(df, target = 'avg_glucose_level', session_id=153, log_experiment=True, experiment_name='health1')
Description Value
0 session_id 153
1 Target avg_glucose_level
2 Original Data (43400, 12)
3 Missing Values True
4 Numeric Features 3
5 Categorical Features 8
6 Ordinal Features False
7 High Cardinality Features False
8 High Cardinality Method None
9 Transformed Train Set (30379, 19)
10 Transformed Test Set (13021, 19)
11 Shuffle Train-Test True
12 Stratify Train-Test False
13 Fold Generator KFold
14 Fold Number 10
15 CPU Jobs -1
16 Use GPU False
17 Log Experiment True
18 Experiment Name health1
19 USI 2590
20 Imputation Type simple
21 Iterative Imputation Iteration None
22 Numeric Imputer mean
23 Iterative Imputation Numeric Model None
24 Categorical Imputer constant
25 Iterative Imputation Categorical Model None
26 Unknown Categoricals Handling least_frequent
27 Normalize False
28 Normalize Method None
29 Transformation False
30 Transformation Method None
31 PCA False
32 PCA Method None
33 PCA Components None
34 Ignore Low Variance False
35 Combine Rare Levels False
36 Rare Level Threshold None
37 Numeric Binning False
38 Remove Outliers False
39 Outliers Threshold None
40 Remove Multicollinearity False
41 Multicollinearity Threshold None
42 Clustering False
43 Clustering Iteration None
44 Polynomial Features False
45 Polynomial Degree None
46 Trignometry Features False
47 Polynomial Threshold None
48 Group Features False
49 Feature Selection False
50 Feature Selection Method classic
51 Features Selection Threshold None
52 Feature Interaction False
53 Feature Ratio False
54 Interaction Threshold None
55 Transform Target False
56 Transform Target Method box-cox

3. Compare Baseline 

best_model = compare_models(fold=5)
Model MAE MSE RMSE R2 RMSLE MAPE TT (Sec)
gbr Gradient Boosting Regressor 29.9120 1671.5862 40.8836 0.1111 0.3360 0.2897 0.7480
lightgbm Light Gradient Boosting Machine 29.8101 1680.0991 40.9877 0.1065 0.3369 0.2889 0.1120
ridge Ridge Regression 30.3617 1705.3918 41.2947 0.0931 0.3394 0.2936 0.0400
lar Least Angle Regression 30.3618 1705.3961 41.2948 0.0931 0.3394 0.2936 0.0300
br Bayesian Ridge 30.3665 1705.4186 41.2951 0.0931 0.3394 0.2936 0.0400
lr Linear Regression 30.3618 1705.3961 41.2948 0.0931 0.3394 0.2936 0.4600
en Elastic Net 30.7817 1744.7248 41.7681 0.0722 0.3428 0.2971 0.0340
lasso Lasso Regression 30.8121 1747.2784 41.7986 0.0709 0.3431 0.2974 0.2980
omp Orthogonal Matching Pursuit 30.8859 1772.4020 42.0980 0.0575 0.3440 0.2968 0.0300
rf Random Forest Regressor 31.0628 1783.7249 42.2324 0.0515 0.3507 0.3061 2.4020
llar Lasso Least Angle Regression 31.0430 1880.7945 43.3662 -0.0001 0.3516 0.2960 0.0300
et Extra Trees Regressor 32.0922 1941.3104 44.0589 -0.0324 0.3645 0.3153 1.6780
ada AdaBoost Regressor 36.2249 1942.8968 44.0775 -0.0335 0.3835 0.3864 0.2040
huber Huber Regressor 30.5788 1947.3049 44.1187 -0.0359 0.3591 0.2734 0.2000
knn K Neighbors Regressor 32.9997 2098.5319 45.8082 -0.1160 0.3778 0.3161 0.0760
dt Decision Tree Regressor 40.5230 3465.8436 58.8670 -0.8435 0.4733 0.3972 0.0820
par Passive Aggressive Regressor 61.6977 7002.6674 77.1516 -2.6714 0.7598 0.6310 0.0500

4. Create Model 

gbr = create_model('gbr')
MAE MSE RMSE R2 RMSLE MAPE
0 29.4483 1605.5183 40.0689 0.1065 0.3327 0.2893
1 30.1403 1696.7591 41.1917 0.1159 0.3351 0.2869
2 29.9468 1650.7967 40.6300 0.1151 0.3359 0.2924
3 30.9916 1797.2590 42.3941 0.1164 0.3429 0.2943
4 29.7740 1654.5201 40.6758 0.1206 0.3339 0.2886
5 29.4301 1644.3258 40.5503 0.1098 0.3356 0.2880
6 29.8422 1649.4613 40.6136 0.1063 0.3369 0.2940
7 30.1508 1683.2950 41.0280 0.1142 0.3346 0.2861
8 29.8594 1681.1659 41.0020 0.1008 0.3383 0.2920
9 29.4870 1635.5078 40.4414 0.1104 0.3323 0.2853
Mean 29.9070 1669.8609 40.8596 0.1116 0.3358 0.2897
SD 0.4393 49.2434 0.5964 0.0056 0.0029 0.0031 
import numpy as np
gbrs = [create_model('gbr', learning_rate=i) for i in np.arange(0.1,1,0.1)]
MAE MSE RMSE R2 RMSLE MAPE
0 30.0877 1704.1490 41.2813 0.0517 0.3429 0.2952
1 31.1447 1869.6863 43.2399 0.0258 0.3573 0.2956
2 30.5247 1759.6697 41.9484 0.0568 0.3475 0.2990
3 31.8691 1938.1533 44.0245 0.0471 0.3581 0.3021
4 30.6442 1808.0313 42.5210 0.0390 0.3494 0.2982
5 30.0122 1742.8764 41.7478 0.0564 0.3456 0.2943
6 30.5092 1785.0937 42.2504 0.0328 0.3503 0.3013
7 30.6449 1770.6888 42.0796 0.0682 0.3457 0.2922
8 30.6724 1813.1471 42.5811 0.0302 0.3533 0.2999
9 30.3654 1773.0299 42.1074 0.0356 0.3480 0.2961
Mean 30.6475 1796.4526 42.3781 0.0444 0.3498 0.2974
SD 0.5070 63.1980 0.7402 0.0131 0.0048 0.0030 
print(len(gbrs))

9

5. Tune Hyperparameters 

tuned_gbr = tune_model(gbr, n_iter=50, optimize = 'RMSE')
MAE MSE RMSE R2 RMSLE MAPE
0 29.4002 1615.6882 40.1956 0.1009 0.3328 0.2882
1 30.0942 1699.4660 41.2246 0.1145 0.3346 0.2859
2 29.9120 1656.1440 40.6957 0.1123 0.3358 0.2914
3 30.9406 1808.0163 42.5208 0.1111 0.3430 0.2928
4 29.8471 1665.1412 40.8061 0.1149 0.3346 0.2890
5 29.4329 1643.2520 40.5370 0.1104 0.3351 0.2880
6 29.6530 1636.3363 40.4517 0.1134 0.3355 0.2918
7 30.1129 1686.1293 41.0625 0.1127 0.3343 0.2850
8 29.8222 1683.4829 41.0303 0.0996 0.3380 0.2913
9 29.3902 1638.9886 40.4844 0.1085 0.3319 0.2840
Mean 29.8605 1673.2645 40.9009 0.1098 0.3356 0.2887
SD 0.4414 51.2132 0.6187 0.0051 0.0029 0.0029 
tuned_gbr

GradientBoostingRegressor(alpha=0.9, ccp_alpha=0.0, criterion='friedman_mse',
                          init=None, learning_rate=0.01, loss='ls', max_depth=8,
                          max_features='sqrt', max_leaf_nodes=None,
                          min_impurity_decrease=0.2, min_impurity_split=None,
                          min_samples_leaf=5, min_samples_split=10,
                          min_weight_fraction_leaf=0.0, n_estimators=280,
                          n_iter_no_change=None, presort='deprecated',
                          random_state=153, subsample=0.65, tol=0.0001,
                          validation_fraction=0.1, verbose=0, warm_start=False)

6. Ensemble Model 

dt = create_model('dt')
MAE MSE RMSE R2 RMSLE MAPE
0 38.8967 3256.2671 57.0637 -0.8121 0.4602 0.3865
1 40.8905 3506.9767 59.2197 -0.8272 0.4737 0.3927
2 39.8432 3339.3332 57.7870 -0.7899 0.4682 0.3920
3 41.0738 3547.0195 59.5569 -0.7439 0.4741 0.3947
4 41.0748 3545.6724 59.5455 -0.8846 0.4780 0.4037
5 40.8333 3518.6152 59.3179 -0.9049 0.4776 0.4027
6 40.5197 3506.7571 59.2179 -0.9000 0.4754 0.4060
7 40.0449 3418.0876 58.4644 -0.7987 0.4696 0.3896
8 40.7124 3463.5019 58.8515 -0.8525 0.4754 0.4031
9 39.4837 3352.3464 57.8995 -0.8234 0.4671 0.3819
Mean 40.3373 3445.4577 58.6924 -0.8337 0.4719 0.3953
SD 0.7033 94.8957 0.8124 0.0492 0.0053 0.0078 
bagged_dt = ensemble_model(dt, n_estimators=50)
MAE MSE RMSE R2 RMSLE MAPE
0 30.4101 1705.5270 41.2980 0.0509 0.3458 0.3031
1 31.1895 1815.9909 42.6144 0.0538 0.3495 0.2995
2 31.1459 1782.9179 42.2246 0.0443 0.3522 0.3090
3 32.1774 1926.5104 43.8920 0.0528 0.3581 0.3104
4 31.4072 1807.7021 42.5171 0.0392 0.3534 0.3106
5 30.8286 1787.9650 42.2843 0.0320 0.3533 0.3073
6 31.0847 1802.8904 42.4605 0.0232 0.3557 0.3122
7 31.2760 1797.0518 42.3916 0.0544 0.3507 0.3029
8 31.1355 1795.7085 42.3758 0.0396 0.3535 0.3108
9 30.7127 1760.7991 41.9619 0.0423 0.3486 0.3020
Mean 31.1368 1798.3063 42.4020 0.0432 0.3521 0.3068
SD 0.4454 52.1778 0.6116 0.0097 0.0034 0.0043 
boosted_dt = ensemble_model(dt, method = 'Boosting')
MAE MSE RMSE R2 RMSLE MAPE
0 32.0837 2316.4857 48.1299 -0.2891 0.3905 0.3044
1 32.8474 2379.9324 48.7846 -0.2400 0.3920 0.3000
2 32.5372 2325.1765 48.2201 -0.2463 0.3920 0.3091
3 33.3153 2448.9699 49.4871 -0.2040 0.3948 0.3026
4 32.6835 2316.9897 48.1351 -0.2315 0.3894 0.3049
5 32.3224 2299.7394 47.9556 -0.2450 0.3899 0.3068
6 32.6263 2307.7303 48.0388 -0.2503 0.3927 0.3125
7 32.4531 2287.3343 47.8261 -0.2036 0.3883 0.3019
8 32.1537 2270.1737 47.6463 -0.2142 0.3867 0.3068
9 31.9558 2247.6878 47.4098 -0.2225 0.3851 0.2979
Mean 32.4978 2320.0220 48.1633 -0.2347 0.3901 0.3047
SD 0.3820 54.4179 0.5613 0.0244 0.0028 0.0041

9. Analyze Model 

plot_model(dt)

plot_model(dt, plot = 'error')

plot_model(dt, plot = 'feature')

evaluate_model(dt)
Parameters
ccp_alpha 0.0
criterion mse
max_depth None
max_features None
max_leaf_nodes None
min_impurity_decrease 0.0
min_impurity_split None
min_samples_leaf 1
min_samples_split 2
min_weight_fraction_leaf 0.0
presort deprecated
random_state 153
splitter best

10. Interpret Model 

interpret_model(dt)

interpret_model(dt, plot = 'correlation')

interpret_model(dt, plot = 'reason', observation = 12)
Visualization omitted, Javascript library not loaded!
Have you run `initjs()` in this notebook? If this notebook was from another user you must also trust this notebook (File -> Trust notebook). If you are viewing this notebook on github the Javascript has been stripped for security. If you are using JupyterLab this error is because a JupyterLab extension has not yet been written.

11. AutoML() 

best = automl(optimize = 'MAE')
best

LGBMRegressor(boosting_type='gbdt', class_weight=None, colsample_bytree=1.0,
              importance_type='split', learning_rate=0.1, max_depth=-1,
              min_child_samples=20, min_child_weight=0.001, min_split_gain=0.0,
              n_estimators=100, n_jobs=-1, num_leaves=31, objective=None,
              random_state=153, reg_alpha=0.0, reg_lambda=0.0, silent=True,
              subsample=1.0, subsample_for_bin=200000, subsample_freq=0)

12. Predict Model 

pred_holdouts = predict_model(dt)
pred_holdouts.head()
Model MAE MSE RMSE R2 RMSLE MAPE
0 Decision Tree Regressor 39.4954 3317.2600 57.5957 -0.8368 0.4665 0.3924
id age bmi gender_Male gender_Other hypertension_1 heart_disease_0 ever_married_Yes work_type_Govt_job work_type_Never_worked ... work_type_Self-employed work_type_children Residence_type_Urban smoking_status_formerly smoked smoking_status_never smoked smoking_status_not_available smoking_status_smokes stroke_0 avg_glucose_level Label
0 22489.0 55.0 41.099998 0.0 0.0 0.0 1.0 1.0 0.0 0.0 ... 0.0 0.0 1.0 0.0 0.0 1.0 0.0 1.0 99.110001 109.269997
1 26145.0 17.0 16.200001 0.0 0.0 0.0 1.0 0.0 0.0 0.0 ... 0.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0 83.470001 73.949997
2 34373.0 0.4 20.200001 1.0 0.0 0.0 1.0 0.0 0.0 0.0 ... 0.0 1.0 1.0 0.0 0.0 1.0 0.0 1.0 76.430000 92.430000
3 60599.0 69.0 32.500000 1.0 0.0 0.0 1.0 1.0 0.0 0.0 ... 0.0 0.0 1.0 0.0 0.0 1.0 0.0 1.0 80.370003 86.099998
4 54268.0 35.0 27.200001 1.0 0.0 0.0 1.0 1.0 0.0 0.0 ... 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 151.589996 175.320007

5 rows × 21 columns

new_data = df.copy()
new_data.drop(['avg_glucose_level'], axis=1, inplace=True)
predict_new = predict_model(best, data=new_data)
predict_new.head()
id gender age hypertension heart_disease ever_married work_type Residence_type bmi smoking_status stroke Label
0 30669 Male 3.0 0 0 No children Rural 18.0 NaN 0 91.930029
1 30468 Male 58.0 1 0 Yes Private Urban 39.2 never smoked 0 143.044226
2 16523 Female 8.0 0 0 No Private Urban 17.6 NaN 0 91.227274
3 56543 Female 70.0 0 0 Yes Private Rural 35.9 formerly smoked 0 123.386541
4 46136 Male 14.0 0 0 No Never_worked Rural 19.1 NaN 0 97.715823

13. Save / Load Model 

save_model(best, model_name='best-model')

Transformation Pipeline and Model Succesfully Saved

(Pipeline(memory=None,
          steps=[('dtypes',
                  DataTypes_Auto_infer(categorical_features=[],
                                       display_types=True, features_todrop=[],
                                       id_columns=[], ml_usecase='regression',
                                       numerical_features=[],
                                       target='avg_glucose_level',
                                       time_features=[])),
                 ('imputer',
                  Simple_Imputer(categorical_strategy='not_available',
                                 fill_value_categorical=None,
                                 fill_value_numerical=None,
                                 numeri...
                  LGBMRegressor(boosting_type='gbdt', class_weight=None,
                                colsample_bytree=1.0, importance_type='split',
                                learning_rate=0.1, max_depth=-1,
                                min_child_samples=20, min_child_weight=0.001,
                                min_split_gain=0.0, n_estimators=100, n_jobs=-1,
                                num_leaves=31, objective=None, random_state=153,
                                reg_alpha=0.0, reg_lambda=0.0, silent=True,
                                subsample=1.0, subsample_for_bin=200000,
                                subsample_freq=0)]],
          verbose=False),
 'best-model.pkl')
loaded_bestmodel = load_model('best-model')
print(loaded_bestmodel)

Transformation Pipeline and Model Successfully Loaded
Pipeline(memory=None,
         steps=[('dtypes',
                 DataTypes_Auto_infer(categorical_features=[],
                                      display_types=True, features_todrop=[],
                                      id_columns=[], ml_usecase='regression',
                                      numerical_features=[],
                                      target='avg_glucose_level',
                                      time_features=[])),
                ('imputer',
                 Simple_Imputer(categorical_strategy='not_available',
                                fill_value_categorical=None,
                                fill_value_numerical=None,
                                numeri...
                 LGBMRegressor(boosting_type='gbdt', class_weight=None,
                               colsample_bytree=1.0, importance_type='split',
                               learning_rate=0.1, max_depth=-1,
                               min_child_samples=20, min_child_weight=0.001,
                               min_split_gain=0.0, n_estimators=100, n_jobs=-1,
                               num_leaves=31, objective=None, random_state=153,
                               reg_alpha=0.0, reg_lambda=0.0, silent=True,
                               subsample=1.0, subsample_for_bin=200000,
                               subsample_freq=0)]],
         verbose=False)

from sklearn import set_config
set_config(display='diagram')
loaded_bestmodel[0]

DataTypes_Auto_infer(categorical_features=[], display_types=True,
                     features_todrop=[], id_columns=[], ml_usecase='regression',
                     numerical_features=[], target='avg_glucose_level',
                     time_features=[])
from sklearn import set_config
set_config(display='text')

14. Deploy Model 

#deploy_model(best, model_name = 'best-aws', authentication = {'bucket' : 'pycaret-test'})

15. Get Config / Set Config 

X_train = get_config('X_train')
X_train.head()
id age bmi gender_Male gender_Other hypertension_1 heart_disease_0 ever_married_Yes work_type_Govt_job work_type_Never_worked work_type_Private work_type_Self-employed work_type_children Residence_type_Urban smoking_status_formerly smoked smoking_status_never smoked smoking_status_not_available smoking_status_smokes stroke_0
41231 51687.0 55.0 25.799999 0.0 0.0 0.0 1.0 1.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0
21949 11886.0 38.0 19.500000 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0 0.0 1.0
12409 18294.0 4.0 18.000000 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 1.0 0.0 1.0
9632 35206.0 37.0 29.000000 0.0 0.0 0.0 1.0 1.0 0.0 0.0 1.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0
15044 69458.0 41.0 38.299999 0.0 0.0 0.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0 
get_config('seed')

153
from pycaret.regression import set_config
set_config('seed', 999)

get_config('seed')

999

16. MLFlow UI 

!mlflow ui

[2021-05-29 11:34:03 -0500] [17806] [INFO] Starting gunicorn 20.0.4
[2021-05-29 11:34:03 -0500] [17806] [INFO] Listening at: http://127.0.0.1:5000 (17806)
[2021-05-29 11:34:03 -0500] [17806] [INFO] Using worker: sync
[2021-05-29 11:34:03 -0500] [17808] [INFO] Booting worker with pid: 17808