You work at a gaming startup that has several terabytes of structured data in Cloud Storage. This data includes gameplay time data, user metadata, and game metadata. You want to build a model that…

Question

Options

• ALoad the data in BigQuery. Use BigQuery ML to train an Autoencoder model.
• BLoad the data in BigQuery. Use BigQuery ML to train a matrix factorization model.
• CRead data to a Vertex AI Workbench notebook. Use TensorFlow to train a two-tower model.
• DRead data to a Vertex AI Workbench notebook. Use TensorFlow to train a matrix factorization

• A
  15% (7)
• B
  72% (34)
• C
  9% (4)
• D
  4% (2)

Explanation

The requirement is a game recommendation system with the least amount of coding. Matrix factorization is the standard collaborative filtering algorithm for recommendation systems-it learns latent factors from user-item interaction data (like gameplay time) to predict preferences. BigQuery ML implements matrix factorization natively via simple SQL statements, requiring no Python, no custom training code, and no infrastructure management. Options C and D require writing TensorFlow model code in a Vertex AI Workbench notebook, which involves significantly more coding. Option A (Autoencoder) is primarily used for dimensionality reduction and anomaly detection, not recommendation systems. BigQuery ML's matrix factorization is the correct tool and the least-code path.

No community discussion yet for this question.