QuestionQ78

VerifiedModeling, Analysis, and Outcomes

A data scientist receives an update to a business case involving a machine with thousands of error codes. While reviewing each machine’s logs, the data scientist creates the following summary-statistics profile:

Question Image

Which of the following is the most likely data-design concern for model ingestion?

Explanation

A high-cardinality error-code field with only a few codes present per machine produces a high-dimensional representation containing mostly zero values when encoded as features. That is a sparse matrix. Amazon SageMaker documentation describes factorization machines as designed for high-dimensional sparse datasets.

Learn more

Community Discussion

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