SnowPro Advanced Data Scientist

By Snowflake · Question Mode

QuestionQ30

VerifiedTrain and use machine learning models

A Data Scientist is developing a fraud-detection model for a shopping website with a high volume of daily transactions. For every transaction, the model will do one of three things: provide a binary recommendation, terminate the transaction, or allow the transaction to proceed.

Which function should be used to evaluate the model’s performance?

Explanation

The F-beta score evaluates classification performance as a weighted harmonic mean of precision and recall. Its beta parameter lets fraud detection prioritize recall or precision according to the relative cost of missed fraud and incorrectly flagged legitimate transactions. Snowflake’s fbeta_score supports binary classification by default and also supports multiclass averaging.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!