Explain the difference between lag and lead in time series analysis.

Instruction: Provide definitions for both 'lag' and 'lead', and give an example of how each might be used in analyzing time series data.

Context: This question tests the candidate's understanding of basic time series concepts and their ability to apply these concepts in practical scenarios.

Updated

Example Answer

A lag is an earlier value, such as yesterday's sales used in a forecast made today. A lead is a later value, such as next week's sales aligned with today's record. The sign used by a software shift function can be confusing, so I would check the timestamps rather than rely on the name.

Leads are useful as future targets or in retrospective analysis once the outcomes are observed. They are not automatically usable predictors. If next week's actual sales appear in a feature for today's forecast, that is leakage. A planned holiday next week can be different because it is known today. I would document when each variable becomes available and recreate that boundary in evaluation.

Make it your own

Give one past feature, one future target, and one future calendar fact known at prediction time.

Why this works

It separates time alignment from information availability.

Follow-up to practice

What if an input is published with a two-day delay?

Related Questions