Example: Holdout set¶

---

This example shows when and how to use ATOM's holdout set in an exploration pipeline.

The data used is a variation on the Australian weather dataset from Kaggle. You can download it from here. The goal of this dataset is to predict whether or not it will rain tomorrow training a binary classifier on target RainTomorrow.

Load the data¶

In [1]:

# Import packages
import pandas as pd
from atom import ATOMClassifier

# Import packages import pandas as pd from atom import ATOMClassifier

UserWarning: The pandas version installed (1.5.3) does not match the supported pandas version in Modin (1.5.2). This may cause undesired side effects!

In [2]:

# Load data
X = pd.read_csv("./datasets/weatherAUS.csv")

# Let's have a look
X.head()

# Load data X = pd.read_csv("./datasets/weatherAUS.csv") # Let's have a look X.head()

Out[2]:

	Location	MinTemp	MaxTemp	Rainfall	Evaporation	Sunshine	WindGustDir	WindGustSpeed	WindDir9am	WindDir3pm	...	Humidity9am	Humidity3pm	Pressure9am	Pressure3pm	Cloud9am	Cloud3pm	Temp9am	Temp3pm	RainToday	RainTomorrow
0	MelbourneAirport	18.0	26.9	21.4	7.0	8.9	SSE	41.0	W	SSE	...	95.0	54.0	1019.5	1017.0	8.0	5.0	18.5	26.0	Yes	0
1	Adelaide	17.2	23.4	0.0	NaN	NaN	S	41.0	S	WSW	...	59.0	36.0	1015.7	1015.7	NaN	NaN	17.7	21.9	No	0
2	Cairns	18.6	24.6	7.4	3.0	6.1	SSE	54.0	SSE	SE	...	78.0	57.0	1018.7	1016.6	3.0	3.0	20.8	24.1	Yes	0
3	Portland	13.6	16.8	4.2	1.2	0.0	ESE	39.0	ESE	ESE	...	76.0	74.0	1021.4	1020.5	7.0	8.0	15.6	16.0	Yes	1
4	Walpole	16.4	19.9	0.0	NaN	NaN	SE	44.0	SE	SE	...	78.0	70.0	1019.4	1018.9	NaN	NaN	17.4	18.1	No	0

5 rows × 22 columns

Run the pipeline¶

In [3]:

# Initialize atom specifying a fraction of the dataset for holdout
atom = ATOMClassifier(X, n_rows=0.5, holdout_size=0.2, verbose=2)

# Initialize atom specifying a fraction of the dataset for holdout atom = ATOMClassifier(X, n_rows=0.5, holdout_size=0.2, verbose=2)

<< ================== ATOM ================== >>
Algorithm task: binary classification.

Dataset stats ==================== >>
Shape: (56877, 22)
Train set size: 42658
Test set size: 14219
Holdout set size: 14219
-------------------------------------
Memory: 24.68 MB
Scaled: False
Missing values: 126457 (10.1%)
Categorical features: 5 (23.8%)
Duplicate samples: 15 (0.0%)

In [4]:

# The test and holdout fractions are split after subsampling the dataset
# Also note that the holdout data set is not a part of atom's dataset
print("Length loaded data:", len(X))
print("Length dataset + holdout:", len(atom.dataset) + len(atom.holdout))

# The test and holdout fractions are split after subsampling the dataset # Also note that the holdout data set is not a part of atom's dataset print("Length loaded data:", len(X)) print("Length dataset + holdout:", len(atom.dataset) + len(atom.holdout))

Length loaded data: 142193
Length dataset + holdout: 71096

In [5]:

atom.impute()
atom.encode()

atom.impute() atom.encode()

Fitting Imputer...
Imputing missing values...
 --> Dropping 246 samples due to missing values in feature MinTemp.
 --> Dropping 136 samples due to missing values in feature MaxTemp.
 --> Dropping 544 samples due to missing values in feature Rainfall.
 --> Dropping 24464 samples due to missing values in feature Evaporation.
 --> Dropping 27155 samples due to missing values in feature Sunshine.
 --> Dropping 3642 samples due to missing values in feature WindGustDir.
 --> Dropping 3616 samples due to missing values in feature WindGustSpeed.
 --> Dropping 4027 samples due to missing values in feature WindDir9am.
 --> Dropping 1491 samples due to missing values in feature WindDir3pm.
 --> Dropping 536 samples due to missing values in feature WindSpeed9am.
 --> Dropping 1035 samples due to missing values in feature WindSpeed3pm.
 --> Dropping 690 samples due to missing values in feature Humidity9am.
 --> Dropping 1419 samples due to missing values in feature Humidity3pm.
 --> Dropping 5566 samples due to missing values in feature Pressure9am.
 --> Dropping 5548 samples due to missing values in feature Pressure3pm.
 --> Dropping 21527 samples due to missing values in feature Cloud9am.
 --> Dropping 22832 samples due to missing values in feature Cloud3pm.
 --> Dropping 373 samples due to missing values in feature Temp9am.
 --> Dropping 1066 samples due to missing values in feature Temp3pm.
 --> Dropping 544 samples due to missing values in feature RainToday.
Fitting Encoder...
Encoding categorical columns...
 --> LeaveOneOut-encoding feature Location. Contains 26 classes.
 --> LeaveOneOut-encoding feature WindGustDir. Contains 16 classes.
 --> LeaveOneOut-encoding feature WindDir9am. Contains 16 classes.
 --> LeaveOneOut-encoding feature WindDir3pm. Contains 16 classes.
 --> Ordinal-encoding feature RainToday. Contains 2 classes.

In [6]:

# Unlike train and test, the holdout data set is not transformed until used for predictions
atom.holdout

# Unlike train and test, the holdout data set is not transformed until used for predictions atom.holdout

Out[6]:

	Location	MinTemp	MaxTemp	Rainfall	Evaporation	Sunshine	WindGustDir	WindGustSpeed	WindDir9am	WindDir3pm	...	Humidity9am	Humidity3pm	Pressure9am	Pressure3pm	Cloud9am	Cloud3pm	Temp9am	Temp3pm	RainToday	RainTomorrow
0	PearceRAAF	13.3	25.6	0.0	NaN	5.8	WNW	52.0	WNW	WNW	...	58.0	51.0	1010.8	1009.6	7.0	7.0	21.6	24.8	No	1
1	Darwin	25.2	30.2	4.2	3.4	2.1	WNW	76.0	WNW	WNW	...	83.0	85.0	1005.9	1003.4	7.0	7.0	28.6	27.4	Yes	1
2	Portland	8.1	15.1	4.4	3.8	7.5	W	52.0	SW	SSW	...	66.0	60.0	1013.6	1017.4	8.0	7.0	13.4	13.2	Yes	0
3	Perth	10.5	22.7	0.0	2.4	9.2	WNW	26.0	NNE	NW	...	86.0	68.0	1016.2	1014.3	1.0	3.0	15.8	20.8	No	1
4	MountGinini	14.7	24.1	102.2	NaN	NaN	NW	52.0	SW	W	...	100.0	78.0	NaN	NaN	NaN	NaN	16.5	21.0	Yes	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
14214	Walpole	10.3	16.1	2.4	NaN	NaN	NW	20.0	NW	SSE	...	97.0	78.0	1024.4	1022.9	NaN	NaN	12.3	14.7	Yes	0
14215	AliceSprings	24.8	29.0	0.2	9.4	0.0	NNW	31.0	NNE	SE	...	69.0	56.0	1009.4	1007.1	8.0	8.0	26.1	27.1	No	0
14216	Darwin	26.2	34.4	0.0	4.4	9.0	NW	33.0	E	WNW	...	69.0	65.0	1009.5	1006.0	3.0	6.0	30.4	32.0	No	0
14217	Cairns	23.9	28.4	2.6	2.6	0.6	ENE	28.0	NE	NE	...	72.0	78.0	1014.4	1011.8	7.0	8.0	27.5	27.0	Yes	1
14218	Tuggeranong	9.5	23.6	0.0	NaN	NaN	W	30.0	NaN	W	...	68.0	40.0	1010.9	1008.4	NaN	NaN	14.9	22.6	No	1

14219 rows × 22 columns

In [7]:

atom.run(models=["GNB", "LR", "RF"])

atom.run(models=["GNB", "LR", "RF"])

Training ========================= >>
Models: GNB, LR, RF
Metric: f1


Results for GaussianNB:
Fit ---------------------------------------------
Train evaluation --> f1: 0.6073
Test evaluation --> f1: 0.6164
Time elapsed: 0.044s
-------------------------------------------------
Total time: 0.044s


Results for LogisticRegression:
Fit ---------------------------------------------
Train evaluation --> f1: 0.6261
Test evaluation --> f1: 0.6333
Time elapsed: 0.134s
-------------------------------------------------
Total time: 0.134s


Results for RandomForest:
Fit ---------------------------------------------
Train evaluation --> f1: 0.9999
Test evaluation --> f1: 0.6163
Time elapsed: 3.344s
-------------------------------------------------
Total time: 3.344s


Final results ==================== >>
Total time: 3.527s
-------------------------------------
GaussianNB         --> f1: 0.6164
LogisticRegression --> f1: 0.6333 !
RandomForest       --> f1: 0.6163 ~

In [8]:

atom.plot_prc()

atom.plot_prc()

In [9]:

# Based on the results on the test set, we select the best model for further tuning
atom.run("lr_tuned", n_trials=10)

# Based on the results on the test set, we select the best model for further tuning atom.run("lr_tuned", n_trials=10)

Training ========================= >>
Models: LR_tuned
Metric: f1


Running hyperparameter tuning for LogisticRegression...
| trial | penalty |       C |  solver | max_iter | l1_ratio |      f1 | best_f1 | time_trial | time_ht |    state |
| ----- | ------- | ------- | ------- | -------- | -------- | ------- | ------- | ---------- | ------- | -------- |
| 0     |    None |     --- | newto.. |      990 |      --- |  0.6325 |  0.6325 |     0.623s |  0.623s | COMPLETE |
| 1     |      l2 |  0.0096 |     sag |      330 |      --- |  0.6076 |  0.6325 |     0.582s |  1.205s | COMPLETE |
| 2     |    None |     --- | newto.. |      490 |      --- |  0.6099 |  0.6325 |     0.567s |  1.772s | COMPLETE |
| 3     |      l2 |  0.0023 | newto.. |      940 |      --- |  0.5731 |  0.6325 |     0.531s |  2.302s | COMPLETE |
| 4     |      l2 | 17.8193 |   lbfgs |      860 |      --- |  0.6297 |  0.6325 |     0.504s |  2.806s | COMPLETE |
| 5     |      l2 | 20.6114 | libli.. |      650 |      --- |  0.6346 |  0.6346 |     0.516s |  3.323s | COMPLETE |
| 6     |      l2 | 74.6411 |     sag |      810 |      --- |  0.6269 |  0.6346 |     0.592s |  3.914s | COMPLETE |
| 7     |      l2 | 36.7273 |   lbfgs |      290 |      --- |  0.6361 |  0.6361 |     0.546s |  4.461s | COMPLETE |
| 8     |    None |     --- | newto.. |      480 |      --- |  0.6328 |  0.6361 |     0.582s |  5.042s | COMPLETE |
| 9     |      l2 |    0.02 |   lbfgs |      760 |      --- |  0.6067 |  0.6361 |     0.506s |  5.549s | COMPLETE |
Hyperparameter tuning ---------------------------
Best trial --> 7
Best parameters:
 --> penalty: l2
 --> C: 36.7273
 --> solver: lbfgs
 --> max_iter: 290
Best evaluation --> f1: 0.6361
Time elapsed: 5.549s
Fit ---------------------------------------------
Train evaluation --> f1: 0.6264
Test evaluation --> f1: 0.6343
Time elapsed: 0.137s
-------------------------------------------------
Total time: 5.686s


Final results ==================== >>
Total time: 5.713s
-------------------------------------
LogisticRegression --> f1: 0.6343

Analyze the results¶

We already used the test set to choose the best model for futher tuning, so this set is no longer truly independent. Although it may not be directly visible in the results, using the test set now to evaluate the tuned LR model would be a mistake, since it carries a bias. For this reason, we have set apart an extra, indepedent set to validate the final model: the holdout set. If we are not going to use the test set for validation, we might as well use it to train the model and so optimize the use of the available data. Use the full_train method for this.

In [10]:

# Re-train the model on the full dataset (train + test) 
atom.lr_tuned.full_train()

# Re-train the model on the full dataset (train + test) atom.lr_tuned.full_train()

Fit ---------------------------------------------
Train evaluation --> f1: 0.626
Test evaluation --> f1: 0.6355
Time elapsed: 0.270s

In [11]:

# Evaluate on the holdout set
atom.lr_tuned.evaluate(dataset="holdout")

# Evaluate on the holdout set atom.lr_tuned.evaluate(dataset="holdout")

Out[11]:

accuracy             0.8550
average_precision    0.7310
balanced_accuracy    0.7396
f1                   0.6185
jaccard              0.4477
matthews_corrcoef    0.5422
precision            0.7358
recall               0.5334
roc_auc              0.8918
Name: LR_tuned, dtype: float64

In [12]:

atom.lr_tuned.plot_prc(dataset="holdout", legend="upper right")

atom.lr_tuned.plot_prc(dataset="holdout", legend="upper right")