Drift Detection for Model Targets

Why Perform Drift Detection for Model Targets

The performance of a machine learning model can be affected if the distribution of targets changes. The target distribution can change both because of data drift but also because of label shift.

Moreover a change in the target distribution may mean that business assumptions on which the model is used may need to be revisited.

Just The Code

If you just want the code to experiment yourself within a Jupyter Notebook, here you go:

>>> import nannyml as nml
>>> import pandas as pd
>>> from IPython.display import display
>>> reference, analysis, analysis_targets = nml.load_synthetic_binary_classification_dataset()
>>> metadata = nml.extract_metadata(data = reference, model_name='wfh_predictor', model_type='classification_binary', exclude_columns=['identifier'])
>>> metadata.target_column_name = 'work_home_actual'
>>> display(reference.head(3))

>>> data = pd.concat([reference, analysis.set_index('identifier').join(analysis_targets.set_index('identifier'), on='identifier', rsuffix='_r')], ignore_index=True).reset_index(drop=True)
>>> display(data.loc[data['partition'] == 'analysis'].head(3))

>>> target_distribution_calculator = nml.TargetDistributionCalculator(model_metadata=metadata, chunk_size=5000)
>>> target_distribution_calculator = target_distribution_calculator.fit(reference_data=reference)

>>> target_distribution = target_distribution_calculator.calculate(data)
>>> display(target_distribution.data.head(3))

>>> fig = target_distribution.plot(kind='distribution', distribution='metric')
>>> fig.show()

>>> fig = target_distribution.plot(kind='distribution', distribution='statistical')
>>> fig.show()

Walkthrough on Drift Detection for Model Targets

Let’s start by loading some synthetic data provided by the NannyML package.

>>> import nannyml as nml
>>> import pandas as pd
>>> from IPython.display import display
>>> reference, analysis, analysis_targets = nml.load_synthetic_binary_classification_dataset()
>>> metadata = nml.extract_metadata(data = reference, model_name='wfh_predictor', model_type='classification_binary', exclude_columns=['identifier'])
>>> metadata.target_column_name = 'work_home_actual'
>>> display(reference.head(3))

	distance_from_office	salary_range	gas_price_per_litre	public_transportation_cost	wfh_prev_workday	workday	tenure	identifier	work_home_actual	timestamp	y_pred_proba	partition	y_pred
0	5.96225	40K - 60K €	2.11948	8.56806	False	Friday	0.212653	0	1	2014-05-09 22:27:20	0.99	reference	1
1	0.535872	40K - 60K €	2.3572	5.42538	True	Tuesday	4.92755	1	0	2014-05-09 22:59:32	0.07	reference	0
2	1.96952	40K - 60K €	2.36685	8.24716	False	Monday	0.520817	2	1	2014-05-09 23:48:25	1	reference	1

NannyML uses TargetDistributionCalculator in order to monitor drift in Target distribution. It can calculate the mean occurrence of positive events for binary classification problems. It can also calculates the chi-squared statistic, from the 2 sample Chi Squared test, of the target values for each chunk which is available for both binary and multiclass classification problems.

In order to calculate target drift, the target values must be available. Let’s manually add the target data to the analysis data first.

Note

The Target Drift detection process can handle missing target values across all data periods.

>>> data = pd.concat([reference, analysis.set_index('identifier').join(analysis_targets.set_index('identifier'), on='identifier', rsuffix='_r')], ignore_index=True).reset_index(drop=True)
>>> display(data.loc[data['partition'] == 'analysis'].head(3))

	distance_from_office	salary_range	gas_price_per_litre	public_transportation_cost	wfh_prev_workday	workday	tenure	identifier	work_home_actual	timestamp	y_pred_proba	partition	y_pred
50000	0.527691	0 - 20K €	1.8	8.96072	False	Tuesday	4.22463	nan	1	2017-08-31 04:20:00	0.99	analysis	1
50001	8.48513	20K - 40K €	2.22207	8.76879	False	Friday	4.9631	nan	1	2017-08-31 05:16:16	0.98	analysis	1
50002	2.07388	40K - 60K €	2.31008	8.64998	True	Friday	4.58895	nan	1	2017-08-31 05:56:44	0.98	analysis	1

Now that the data is in place we’ll create a new TargetDistributionCalculator instantiating it with appropriate parameters. Afterwards, the fit() method gets called on the reference period which represent an accepted target distribution against which we will compare data from the analysis period. Then the calculate() method gets called to calculate the target drift results on the data provided to it. We use the previously assembled data as an argument.

>>> target_distribution_calculator = nml.TargetDistributionCalculator(model_metadata=metadata, chunk_size=5000)
>>> target_distribution_calculator = target_distribution_calculator.fit(reference_data=reference)
>>> target_distribution = target_distribution_calculator.calculate(data)
>>> display(target_distribution.data.head(3))

	key	start_index	end_index	start_date	end_date	partition	metric_target_drift	statistical_target_drift	p_value	thresholds	alert	significant
0	[0:4999]	0	4999	2014-05-09 22:27:20	2014-09-09 08:18:27	reference	0.4944	0.467363	0.494203	0.05	False	False
1	[5000:9999]	5000	9999	2014-09-09 09:13:35	2015-01-09 00:02:51	reference	0.493	0.76111	0.382981	0.05	False	False
2	[10000:14999]	10000	14999	2015-01-09 00:04:43	2015-05-09 15:54:26	reference	0.505	0.512656	0.473991	0.05	False	False

The results can be easily plotted by using the plot() method.

>>> fig = target_distribution.plot(kind='distribution', distribution='metric')
>>> fig.show()

Note that a dashed line, instead of a solid line, will be used for chunks that have missing target values.

>>> fig = target_distribution.plot(kind='distribution', distribution='statistical')
>>> fig.show()

Insights and Follow Ups

Looking at the results we see that we have a false alert on the first chunk of the analysis data. This can happen when the statistical tests consider significant a small change in the distribution of a variable in the chunks. But because the change is small it is usually not significant from a model monitoring perspective.

The Monitoring Realized Performance functionality of NannyML can can add context to the target drift results showing whether there are associated performance changes.