1

Missing values are very common in large time series data. How should the missing values of a time series be estimated? Is interpolation useful or I need to forecast them from the past values?

JRK
  • 603
  • 5
  • 15
  • 3
    Searching the archive would give you multiple similar questions http://stats.stackexchange.com/search?q=[time-series]+missing+data – Tim Mar 24 '15 at 16:51
  • 2
    Interesting question. You may get more and better answers if you could add some details on what kind of time series you have (just as ["short time series" can be very different things to different people](http://stats.stackexchange.com/questions/135061/best-method-for-short-time-series/135146#comment257352_135080), "large time series data" can be many different things) and where the "missingness" comes from. – Stephan Kolassa Mar 24 '15 at 16:52
  • Missing values are in the middle of a long time series. – JRK Mar 24 '15 at 16:59
  • 1
    @Rudro88 Stephen meant rather *how large* is your series and *how* did the missing values appeared (are they missing completely at random, at random, not at random etc.). – Tim Mar 24 '15 at 17:19
  • I have data of daily basis of 15 years. Missing values are at completely random manner. – JRK Mar 24 '15 at 17:44
  • As with some of the linked questions, the answer strongly depends on what you want to do with the time series. There are several situations, in which you would not need to do anything at all. – Wrzlprmft Mar 24 '15 at 21:40

0 Answers0