Datasets Size Comparison¶
[1]:
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from deepchecks import Dataset
from deepchecks.checks.methodology import DatasetsSizeComparison
[2]:
df = pd.DataFrame(np.random.randn(1000, 3), columns=['x1', 'x2', 'x3'])
df['label'] = df['x2'] + 0.1 * df['x1']
train, test = train_test_split(df, test_size=0.4)
train = Dataset(train, features=['x1', 'x2', 'x3'], label='label')
test = Dataset(test, features=['x1', 'x2', 'x3'], label='label')
check_instance = (
DatasetsSizeComparison()
.add_condition_train_dataset_not_smaller_than_test()
.add_condition_test_size_not_smaller_than(100)
.add_condition_test_train_size_ratio_not_smaller_than(0.2)
)
[3]:
check_instance.run(train, test)
Datasets Size Comparison
Verify test dataset size comparing it to the train dataset size. Read More...
Conditions Summary
| Status | Condition | More Info |
|---|---|---|
✓ |
Train dataset is not smaller than test dataset | |
✓ |
Test dataset size is not smaller than 100 | |
✓ |
Test-Train size ratio is not smaller than 0.2 |
Additional Outputs
| Train | Test | |
|---|---|---|
| Size | 600 | 400 |