PyCaret 2 Classification Example¶

This notebook is created using PyCaret 2.0. Last updated : 28-07-2020

In [1]:

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

pycaret-nightly-0.39

1. Data Repository¶

In [2]:

from pycaret.datasets import get_data
index = get_data('index')

	Dataset	Data Types	Default Task	Target Variable	# Instances	# Attributes	Missing Values
0	anomaly	Multivariate	Anomaly Detection	None	1000	10	N
1	france	Multivariate	Association Rule Mining	InvoiceNo, Description	8557	8	N
2	germany	Multivariate	Association Rule Mining	InvoiceNo, Description	9495	8	N
3	bank	Multivariate	Classification (Binary)	deposit	45211	17	N
4	blood	Multivariate	Classification (Binary)	Class	748	5	N
5	cancer	Multivariate	Classification (Binary)	Class	683	10	N
6	credit	Multivariate	Classification (Binary)	default	24000	24	N
7	diabetes	Multivariate	Classification (Binary)	Class variable	768	9	N
8	electrical_grid	Multivariate	Classification (Binary)	stabf	10000	14	N
9	employee	Multivariate	Classification (Binary)	left	14999	10	N
10	heart	Multivariate	Classification (Binary)	DEATH	200	16	N
11	heart_disease	Multivariate	Classification (Binary)	Disease	270	14	N
12	hepatitis	Multivariate	Classification (Binary)	Class	154	32	Y
13	income	Multivariate	Classification (Binary)	income >50K	32561	14	Y
14	juice	Multivariate	Classification (Binary)	Purchase	1070	15	N
15	nba	Multivariate	Classification (Binary)	TARGET_5Yrs	1340	21	N
16	wine	Multivariate	Classification (Binary)	type	6498	13	N
17	telescope	Multivariate	Classification (Binary)	Class	19020	11	N
18	glass	Multivariate	Classification (Multiclass)	Type	214	10	N
19	iris	Multivariate	Classification (Multiclass)	species	150	5	N
20	poker	Multivariate	Classification (Multiclass)	CLASS	100000	11	N
21	questions	Multivariate	Classification (Multiclass)	Next_Question	499	4	N
22	satellite	Multivariate	Classification (Multiclass)	Class	6435	37	N
23	asia_gdp	Multivariate	Clustering	None	40	11	N
24	elections	Multivariate	Clustering	None	3195	54	Y
25	facebook	Multivariate	Clustering	None	7050	12	N
26	ipl	Multivariate	Clustering	None	153	25	N
27	jewellery	Multivariate	Clustering	None	505	4	N
28	mice	Multivariate	Clustering	None	1080	82	Y
29	migration	Multivariate	Clustering	None	233	12	N
30	perfume	Multivariate	Clustering	None	20	29	N
31	pokemon	Multivariate	Clustering	None	800	13	Y
32	population	Multivariate	Clustering	None	255	56	Y
33	public_health	Multivariate	Clustering	None	224	21	N
34	seeds	Multivariate	Clustering	None	210	7	N
35	wholesale	Multivariate	Clustering	None	440	8	N
36	tweets	Text	NLP	tweet	8594	2	N
37	amazon	Text	NLP / Classification	reviewText	20000	2	N
38	kiva	Text	NLP / Classification	en	6818	7	N
39	spx	Text	NLP / Regression	text	874	4	N
40	wikipedia	Text	NLP / Classification	Text	500	3	N
41	automobile	Multivariate	Regression	price	202	26	Y
42	bike	Multivariate	Regression	cnt	17379	15	N
43	boston	Multivariate	Regression	medv	506	14	N
44	concrete	Multivariate	Regression	strength	1030	9	N
45	diamond	Multivariate	Regression	Price	6000	8	N
46	energy	Multivariate	Regression	Heating Load / Cooling Load	768	10	N
47	forest	Multivariate	Regression	area	517	13	N
48	gold	Multivariate	Regression	Gold_T+22	2558	121	N
49	house	Multivariate	Regression	SalePrice	1461	81	Y
50	insurance	Multivariate	Regression	charges	1338	7	N
51	parkinsons	Multivariate	Regression	PPE	5875	22	N
52	traffic	Multivariate	Regression	traffic_volume	48204	8	N

In [4]:

data = get_data('juice')

	Id	Purchase	WeekofPurchase	StoreID	PriceCH	PriceMM	DiscCH	DiscMM	SpecialMM	LoyalCH	SalePriceMM	SalePriceCH	PriceDiff	Store7	PctDiscMM	PctDiscCH	ListPriceDiff	STORE
0	1	CH	237	1	1.75	1.99	0.00	0.0	0	0.500000	1.99	1.75	0.24	No	0.000000	0.000000	0.24	1
1	2	CH	239	1	1.75	1.99	0.00	0.3	1	0.600000	1.69	1.75	-0.06	No	0.150754	0.000000	0.24	1
2	3	CH	245	1	1.86	2.09	0.17	0.0	0	0.680000	2.09	1.69	0.40	No	0.000000	0.091398	0.23	1
3	4	MM	227	1	1.69	1.69	0.00	0.0	0	0.400000	1.69	1.69	0.00	No	0.000000	0.000000	0.00	1
4	5	CH	228	7	1.69	1.69	0.00	0.0	0	0.956535	1.69	1.69	0.00	Yes	0.000000	0.000000	0.00	0

2. Initialize Setup¶

In [5]:

from pycaret.classification import *
clf1 = setup(data, target = 'Purchase', session_id=123, log_experiment=False, experiment_name='bank1')

Setup Succesfully Completed!

	Description	Value
0	session_id	123
1	Target Type	Binary
2	Label Encoded	CH: 0, MM: 1
3	Original Data	(1070, 19)
4	Missing Values	False
5	Numeric Features	13
6	Categorical Features	5
7	Ordinal Features	False
8	High Cardinality Features	False
9	High Cardinality Method	None
10	Sampled Data	(1070, 19)
11	Transformed Train Set	(748, 28)
12	Transformed Test Set	(322, 28)
13	Numeric Imputer	mean
14	Categorical Imputer	constant
15	Normalize	False
16	Normalize Method	None
17	Transformation	False
18	Transformation Method	None
19	PCA	False
20	PCA Method	None
21	PCA Components	None
22	Ignore Low Variance	False
23	Combine Rare Levels	False
24	Rare Level Threshold	None
25	Numeric Binning	False
26	Remove Outliers	False
27	Outliers Threshold	None
28	Remove Multicollinearity	False
29	Multicollinearity Threshold	None
30	Clustering	False
31	Clustering Iteration	None
32	Polynomial Features	False
33	Polynomial Degree	None
34	Trignometry Features	False
35	Polynomial Threshold	None
36	Group Features	False
37	Feature Selection	False
38	Features Selection Threshold	None
39	Feature Interaction	False
40	Feature Ratio	False
41	Interaction Threshold	None
42	Fix Imbalance	False
43	Fix Imbalance Method	SMOTE

3. Compare Baseline¶

In [6]:

top5 = compare_models(n_select=5)

	Model	Accuracy	AUC	Recall	Prec.	F1	Kappa	MCC	TT (Sec)
0	Logistic Regression	0.8263	0.8959	0.7262	0.8139	0.7644	0.6280	0.6338	0.0429
1	Linear Discriminant Analysis	0.8263	0.8938	0.7536	0.7938	0.7713	0.6317	0.6342	0.0089
2	Ridge Classifier	0.8236	0.0000	0.7499	0.7920	0.7680	0.6262	0.6292	0.0045
3	Ada Boost Classifier	0.8075	0.8637	0.7053	0.7837	0.7398	0.5881	0.5924	0.0780
4	Gradient Boosting Classifier	0.8062	0.8869	0.7363	0.7651	0.7479	0.5909	0.5939	0.1205
5	CatBoost Classifier	0.8049	0.8932	0.7326	0.7629	0.7457	0.5878	0.5899	3.5314
6	Extreme Gradient Boosting	0.7914	0.8716	0.7294	0.7367	0.7309	0.5609	0.5633	0.0675
7	Light Gradient Boosting Machine	0.7861	0.8806	0.7053	0.7393	0.7195	0.5471	0.5497	0.0898
8	Quadratic Discriminant Analysis	0.7621	0.8240	0.6267	0.7397	0.6678	0.4863	0.5000	0.0060
9	Random Forest Classifier	0.7608	0.8397	0.6674	0.7124	0.6848	0.4928	0.4974	0.1142
10	Decision Tree Classifier	0.7594	0.7519	0.6911	0.6970	0.6907	0.4943	0.4975	0.0051
11	Extra Trees Classifier	0.7433	0.8205	0.6708	0.6758	0.6698	0.4605	0.4638	0.1503
12	K Neighbors Classifier	0.7231	0.7683	0.6062	0.6600	0.6287	0.4094	0.4129	0.0054
13	Naive Bayes	0.7140	0.7952	0.7466	0.6100	0.6708	0.4227	0.4301	0.0032
14	SVM - Linear Kernel	0.5267	0.0000	0.4200	0.2409	0.2561	0.0204	0.0299	0.0074

In [7]:

top5_tuned = [tune_model(i) for i in top5]

	Accuracy	AUC	Recall	Prec.	F1	Kappa	MCC
0	0.7467	0.8426	0.6897	0.6667	0.6780	0.4693	0.4695
1	0.8267	0.9475	0.7241	0.8077	0.7636	0.6274	0.6298
2	0.7600	0.8201	0.6552	0.7037	0.6786	0.4875	0.4883
3	0.8133	0.9070	0.8276	0.7273	0.7742	0.6162	0.6200
4	0.7867	0.8973	0.8621	0.6757	0.7576	0.5720	0.5856
5	0.8133	0.8906	0.7241	0.7778	0.7500	0.6014	0.6023
6	0.8133	0.8833	0.8333	0.7353	0.7812	0.6196	0.6233
7	0.8400	0.9222	0.8000	0.8000	0.8000	0.6667	0.6667
8	0.7973	0.8759	0.6897	0.7692	0.7273	0.5667	0.5689
9	0.8514	0.9111	0.7586	0.8462	0.8000	0.6823	0.6849
Mean	0.8049	0.8898	0.7564	0.7509	0.7510	0.5909	0.5939
SD	0.0314	0.0354	0.0673	0.0562	0.0420	0.0659	0.0662

4. Create Model¶

In [ ]:

lr = create_model('lr', fold = 5)

In [ ]:

dt = create_model('dt')

In [ ]:

rf = create_model('rf', fold = 5)

In [ ]:

models()

In [ ]:

models(type='ensemble').index.tolist()

In [ ]:

ensembled_models = compare_models(whitelist = models(type='ensemble').index.tolist(), fold = 3)

5. Tune Hyperparameters¶

In [ ]:

tuned_lr = tune_model(lr)

In [ ]:

tuned_rf = tune_model(rf)

6. Ensemble Model¶

In [ ]:

bagged_dt = ensemble_model(dt)

In [ ]:

boosted_dt = ensemble_model(dt, method = 'Boosting')

7. Blend Models¶

In [ ]:

blender = blend_models(estimator_list = [boosted_dt, bagged_dt], method = 'soft')

8. Stack Models¶

In [ ]:

stacker = stack_models(estimator_list = [boosted_dt,bagged_dt,tuned_rf], meta_model=rf)

9. Analyze Model¶

In [ ]:

plot_model(rf)

In [ ]:

plot_model(rf, plot = 'confusion_matrix')

In [ ]:

plot_model(rf, plot = 'boundary')

In [ ]:

evaluate_model(rf)

10. Interpret Model¶

In [ ]:

catboost = create_model('catboost', cross_validation=False)

In [ ]:

interpret_model(catboost)

In [ ]:

interpret_model(catboost, plot = 'correlation')

In [ ]:

interpret_model(catboost, plot = 'reason', observation = 12)

11. AutoML()¶

In [ ]:

best = automl(optimize = 'Recall')
best

12. Predict Model¶

In [ ]:

pred_holdouts = predict_model(lr)
pred_holdouts.head()

In [ ]:

new_data = data.copy()
new_data.drop(['deposit'], axis=1, inplace=True)
predict_new = predict_model(lr, data=new_data)
predict_new.head()

13. Save / Load Model¶

In [ ]:

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

In [ ]:

loaded_bestmodel = load_model('best-model')
print(loaded_bestmodel)

In [ ]:

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

In [ ]:

from sklearn import set_config
set_config(display='text')

14. Deploy Model¶

In [ ]:

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

15. Get Config / Set Config¶

In [ ]:

X_train = get_config('X_train')
X_train.head()

In [ ]:

get_config('seed')

In [ ]:

from pycaret.classification import set_config
set_config('seed', 999)

In [ ]:

get_config('seed')

16. Get System Logs¶

In [ ]:

get_system_logs()

17. MLFlow UI¶

In [ ]:

!mlflow ui

In [ ]:

# to generate csv file with experiment logs
get_logs()

End¶

Thank you. For more information / tutorials on PyCaret, please visit https://www.pycaret.org