Discuss Google Professional Data Engineer Exam Topic 3 Question 122

You work for a large real estate firm and are preparing 6 TB of home sales data lo be used for machine learning You will use SOL to transform the data and use BigQuery ML lo create a machine learning model. You plan to use the model for predictions against a raw dataset that has not been transformed. How should you set up your workflow in order to prevent skew at prediction time?

AWhen creating your model, use BigQuerys TRANSFORM clause to define preprocessing stops. At prediction time, use BigQuery's ML. EVALUATE clause without specifying any transformations on the raw input data.

BWhen creating your model, use BigQuery's TRANSFORM clause to define preprocessing steps Before requesting predictions, use a saved query to transform your raw input data, and then use ML. EVALUATE

CUse a BigOuery to define your preprocessing logic. When creating your model, use the view as your model training data. At prediction lime, use BigQuery's ML EVALUATE clause without specifying any transformations on the raw input data.

DPreprocess all data using Dataflow. At prediction time, use BigOuery's ML. EVALUATE clause without specifying any further transformations on the input data.

Suggested Answer: A

https://cloud.google.com/bigquery-ml/docs/bigqueryml-transform Using the TRANSFORM clause, you can specify all preprocessing during model creation. The preprocessing is automatically applied during the prediction and evaluation phases of machine learning

Get Premium Professional Data Engineer Questions as Interactive Web-Based Practice Test or PDF Real Exam Format Detailed Explanations Instant Access Get Your Practice Test

24 comments · newest first

Belen 2 days ago

I’m leaning towards A. Using ML.EVALUATE without transformations seems risky.

Dwight 8 days ago

Option D could work too. Dataflow is powerful for preprocessing.

Bobbie 13 days ago

I agree, B is solid. Preprocessing ensures consistency.

Brigette 18 days ago

I think option B makes the most sense. Transforming the raw data first is crucial.

Hyun 23 days ago

I feel like B is the safest choice. It minimizes the risk of skew.

Stephaine 28 days ago

C is interesting. Using a view could simplify the workflow.

Brunilda 1 month ago

I’m leaning towards A. Using ML.EVALUATE without transformations seems efficient.

Vi 1 month ago

Option D sounds solid too. Dataflow can handle large datasets well.

Phil 1 month ago

I agree, B ensures consistency. Preprocessing is key for accuracy.

Coletta 2 months ago

I think option B is the best. Transforming the raw data first makes sense.

Noah 2 months ago

C is interesting, using a view could streamline the process!

Pearlene 2 months ago

Wait, can we really skip transformations at prediction time? Sounds off.

Marshall 2 months ago

A is a bit risky, you need consistent transformations for predictions.

Lachelle 2 months ago

I disagree, D is better for handling large datasets efficiently.

Magnolia 2 months ago

Option B seems solid, saves time on preprocessing later.

Lavonda 3 months ago

C is interesting, but using a view might complicate things later on.

Rosamond 3 months ago

Wait, can you really use raw data without any transformations? That sounds sketchy!

Bonita 3 months ago

A is a bit risky, skipping transformations at prediction time could lead to issues.

Erasmo 3 months ago

I disagree, I think D is more efficient with Dataflow.

Ma 3 months ago

Option B seems like the best choice for consistent preprocessing.

Xenia 4 months ago

I recall a similar question where we had to ensure the model and input data were aligned. I feel like option C could work, but I’m not sure if using a view is the right way to go about it.

Ma 4 months ago

I think option D could be a good choice since it suggests preprocessing all data with Dataflow, but I'm not entirely confident if that’s the best approach for this scenario.

Shelton 4 months ago

I'm a bit unsure about the TRANSFORM clause. Wasn't there a question where we had to decide between using it during training or prediction? I feel like option A might lead to issues since it doesn't transform the raw data at prediction time.

Deandrea 5 months ago

I remember we discussed the importance of consistent preprocessing during our practice sessions. I think option B makes sense because it emphasizes transforming the raw data before predictions.

Google Professional Data Engineer Exam - Topic 3 Question 122 Discussion