Assessing Whether a Time-Series Follows a Random Walk

I demonstrate how to assess 3 characteristics of a random walk process on a set of time-series data. A random walk meanders, has differences that form a random process, and has a standard deviation larger than the standard deviation of the differences. To check the first characteristic, I merely make a time-series plot and look for a general pattern of "points close together in time that are also close together in value". To check if the differences are random, create a column of differences, and make a time-series plot of those differences, and look for randomness. Finally, to check the standard deviations merely find descriptive stats on the original series and the differences and compare standard deviations.

Видео Assessing Whether a Time-Series Follows a Random Walk канала ProfTDub