Answer-first summary for fast verification
Answer: Relative Expression Split
For time series data, the correct splitting mode is Relative Expression Split (C) because it allows splitting based on conditions applied to date/time columns, preserving the temporal order and preventing data leakage. This is crucial for time series data where sequential patterns, trends, and seasonality must be maintained. The community discussion strongly supports C (67% consensus), with highly upvoted comments emphasizing that random splits (D) disrupt time series integrity and can cause information leakage. Options A (Recommender Split) and B (Regular Expression Split) are unsuitable as they are for recommendation systems and text pattern matching, respectively, not time series data.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
You are working with a time series dataset in Azure Machine Learning Studio. You need to split the dataset into training and testing subsets using the Split Data module. Which splitting mode should you use?
A
Recommender Split
B
Regular Expression Split
C
Relative Expression Split
D
Split Rows with the Randomized split parameter set to true
No comments yet.