3

I have a simple question: does any data gathered over time could be treated as a time series? For instance, if I measure CPU load at certain intervals, would that be a time series? Actually, I'm not interested on actual time of the measurement. I could simply mark the points along x-axis as 1, 2, 3, ... -- what is relevant to me is certain values at certain points of time.

Further, could I do time series analysis on the data set if I have irregular observation intervals?

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Barun
  • 257
  • 2
  • 9
  • 3
    Yes to everything, by definition. [Wikipedia](http://en.wikipedia.org/wiki/Time_series), [Search 1](http://lmgtfy.com/?q=Time+series), [Search 2](http://lmgtfy.com/?q=time+series+unequal+intervals). –  Apr 25 '12 at 16:01
  • Also note that there are pros and cons to a time series because of the additional relationship of time: it gives you more information, but it also can tie things together in ways that can mislead you. – Wayne Apr 25 '12 at 17:07
  • 3
    @Procrastinator, please visit http://meta.stackexchange.com/a/15660/166924. – whuber Apr 25 '12 at 17:43
  • 2
    See https://stats.stackexchange.com/questions/126791 (a later thread) for definitions of time series. – whuber Aug 16 '18 at 13:03
  • Possible duplicate of [Is a time series the same as a stochastic process?](https://stats.stackexchange.com/questions/126791/is-a-time-series-the-same-as-a-stochastic-process) – kjetil b halvorsen Sep 08 '19 at 22:40

1 Answers1

1

Partially answered in comments:

Yes to everything, by definition. https://en.wikipedia.org/wiki/Time_series, https://www.google.com/search?q=Time+series, https://www.google.com/search?q=time+series+unequal+intervals. – user10525

Also note that there are pros and cons to a time series because of the additional relationship of time: it gives you more information, but it also can tie things together in ways that can mislead you. – Wayne

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467