Comparing estimated and realized performance
Now that you have seen how performance calculation works, your task is to calculate the realized performance for our tip prediction model for the NYC green taxi dataset.
The reference and analysis set is already loaded and saved in the reference
and analysis
variables.
In addition, results from the DLE algorithm for tip prediction are stored in the estimated_results
variable.
This exercise is part of the course
Monitoring Machine Learning in Python
Exercise instructions
- Specify problem type as
regression
in calculator initialization. - Fit the calculator with reference data and calculate performance for the analysis set.
- Show comparison plot between
realized_results
andestimated_results
usingcompare()
method.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Intialize the calculator
calculator = nannyml.PerformanceCalculator(
y_true='tip_amount',
y_pred='y_pred',
chunk_period='d',
metrics=['mae'],
timestamp_column_name='lpep_pickup_datetime',
problem_type=____)
# Fit the calculator
calculator.fit(____)
realized_results = calculator.____(____)
# Show comparison plot for realized and estimated performance
____.____(____).plot().show()