2
$\begingroup$

I have a data for each minute as shown below

Timestamp cpu-usage 2017-02-14 00:00:00 1.80000 2017-02-14 00:01:00 16.04000 2017-02-14 00:02:00 23.16000 2017-02-14 00:03:00 24.21400 2017-02-14 00:04:00 24.74100 2017-02-14 00:05:00 12.88767

I have created a graph for this data

>plot(data.ts)

enter image description here

I have even plot Correlation plot for this

>acf((log(data.ts)))

enter image description here

>pacf((log(data.ts)))

Now by seeing the above graphs, how can I make sure that my data is stationary or not. And also how to apply model for forecasting. Which model is best for this data. How to get the frequency for this.

$\endgroup$
1
  • $\begingroup$ I need clarification. The graph shows one day worth of data by minute. There are ten peaks from 0:00 to 2:30, so about one peak every 15 minutes. In the ACF graph, I'm seeing labels of 500 and 1000, so is that x axis in seconds? If so, it looks like you need to "seasonally" difference your data by 15 minutes and see if that improves your graphs. $\endgroup$ – zbicyclist Apr 13 '17 at 21:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.