nerdexam
Google

PROFESSIONAL-MACHINE-LEARNING-ENGINEER · Question #82

You need to train a regression model based on a dataset containing 50,000 records that is stored in BigQuery. The data includes a total of 20 categorical and numerical features with a target…

The correct answer is B. Use BQML XGBoost regression to train the model. For a structured dataset in BigQuery requiring a regression model with balanced effort, training time, and performance, BigQuery ML with XGBoost regression is an efficient and effective choice.

Submitted by asante_acc· Apr 18, 2026ML model development

Question

You need to train a regression model based on a dataset containing 50,000 records that is stored in BigQuery. The data includes a total of 20 categorical and numerical features with a target variable that can include negative values. You need to minimize effort and training time while maximizing model performance. What approach should you take to train this regression model?

Options

  • ACreate a custom TensorFlow DNN model
  • BUse BQML XGBoost regression to train the model.
  • CUse AutoML Tables to train the model without early stopping.
  • DUse AutoML Tables to train the model with RMSLE as the optimization objective.

How the community answered

(49 responses)
  • A
    10% (5)
  • B
    82% (40)
  • C
    2% (1)
  • D
    6% (3)

Why each option

For a structured dataset in BigQuery requiring a regression model with balanced effort, training time, and performance, BigQuery ML with XGBoost regression is an efficient and effective choice.

ACreate a custom TensorFlow DNN model

Creating a custom TensorFlow DNN model involves significant effort in architecture design, hyperparameter tuning, and deployment compared to the highly integrated BQML approach.

BUse BQML XGBoost regression to train the model.Correct

BigQuery ML (BQML) allows training models like XGBoost regression directly within BigQuery using SQL, minimizing data movement and effort. XGBoost is a powerful algorithm known for excellent performance on tabular data and naturally handles negative target values.

CUse AutoML Tables to train the model without early stopping.

While AutoML Tables can be used, training without early stopping can lead to overfitting and wasted training time, contradicting the goal of minimizing training time.

DUse AutoML Tables to train the model with RMSLE as the optimization objective.

RMSLE (Root Mean Squared Log Error) is an inappropriate optimization objective when the target variable can include negative values, as the logarithm of negative numbers is undefined.

Concept tested: Google Cloud ML service and model selection for tabular data

Source: https://cloud.google.com/bigquery-ml/docs/introduction

Topics

#BQML#XGBoost#Regression#Model Selection

Community Discussion

No community discussion yet for this question.

Sign up to join the discussion
Full PROFESSIONAL-MACHINE-LEARNING-ENGINEER Practice