Google Professional Machine Learning Engineer
Get started today
Ultimate access to all questions.
Comments
Loading comments...
Ultimate access to all questions.
As a Machine Learning Engineer at a travel company, you've developed models to predict customer vacation patterns. These models have shown that while destinations vary seasonally, there are consistent trends year over year. Your team is now focused on improving model accuracy and efficiency. To achieve this, you need to compare multiple model iterations and their performance metrics across different seasons and years. The solution must support easy comparison, scalability, and cost-effectiveness. Given these requirements, what is the best approach? Choose the two most appropriate options.
A
Utilize Cloud SQL to store performance metrics and run queries to compare different model versions. This approach is cost-effective but may lack the scalability needed for large datasets.
B
For each season and year, create model versions in Vertex AI and use the Evaluate tab to compare their performance. This method provides a user-friendly interface but may not be the most efficient for large-scale comparisons.
C
Leverage Vertex ML Metadata to store performance metrics, using seasons and years as events, facilitating easy comparison across model versions. This approach is scalable and efficient for comparing large datasets.
D
Store each pipeline run's performance metrics in Kubeflow, organizing them by season and year, and compare through the Kubeflow UI. This method is effective for teams already using Kubeflow but may require additional setup for others.
E
Implement a custom solution using BigQuery to store and analyze performance metrics, enabling complex queries and comparisons. This option offers flexibility and scalability but may incur higher costs.