nerdexam
Amazon

MLS-C01 · Question #378

A data scientist uses Amazon SageMaker Data Wrangler to obtain a feature summary from a dataset that the data scientist imported from Amazon S3. The data scientist notices that the prediction power…

The correct answer is A. Target leakage occurred in the imported dataset. A prediction power score of 1 for a feature in Amazon SageMaker Data Wrangler's summary indicates that target leakage has occurred in the dataset.

Exploratory Data Analysis

Question

A data scientist uses Amazon SageMaker Data Wrangler to obtain a feature summary from a dataset that the data scientist imported from Amazon S3. The data scientist notices that the prediction power for a dataset feature has a score of 1. What is the cause of the score?

Options

  • ATarget leakage occurred in the imported dataset.
  • BThe data scientist did not fine-tune the training and validation split.
  • CThe SageMaker Data Wrangler algorithm that the data scientist used did not find an optimal model
  • DThe data scientist did not process the features enough to accurately calculate prediction power.

How the community answered

(41 responses)
  • A
    85% (35)
  • B
    5% (2)
  • C
    7% (3)
  • D
    2% (1)

Why each option

A prediction power score of 1 for a feature in Amazon SageMaker Data Wrangler's summary indicates that target leakage has occurred in the dataset.

ATarget leakage occurred in the imported dataset.Correct

Target leakage happens when the training data includes information that would not be available at the time of prediction, causing a feature to perfectly or near-perfectly predict the target variable. A prediction power score of 1 signals that a feature has an unrealistic, perfect correlation with the target, which is a strong indicator of target leakage.

BThe data scientist did not fine-tune the training and validation split.

An improper training and validation split might affect overall model evaluation but does not directly cause an individual feature's prediction power to be 1 in a feature summary.

CThe SageMaker Data Wrangler algorithm that the data scientist used did not find an optimal model

The SageMaker Data Wrangler quick model's inability to find an optimal model pertains to the model's performance, not an individual feature's intrinsic prediction power score of 1, which typically points to a data issue.

DThe data scientist did not process the features enough to accurately calculate prediction power.

Insufficient feature processing would more likely lead to lower prediction power or difficulty for a model to learn, rather than a perfect prediction power score of 1 for a feature.

Concept tested: Target leakage in feature engineering

Source: https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler-troubleshoot.html

Topics

#Target Leakage#Feature Engineering#SageMaker Data Wrangler#Data Quality

Community Discussion

No community discussion yet for this question.

Full MLS-C01 Practice