missing data in timeseries

anildas

Hi,
I have a  dataset with daily data . but there are few missing days.. how to add those missing samples with previous day values in the final example set for a timeseries forecast model.
is there any direct process/approach available in Rapidminder for such cases?. I have seen few similar queries but its confusing to me, kindly advise.

regards
Anil

jacobcybulski

Got it, I have misread the original post - you are not having missing values in the examples but missing indices in the series. What you need is to use operator "Equalize Numerical Indices", where you'd define your index attribute, use the "range(start,stop) and step size" method with the original start, stop and step values and use interpolation for numerical and date-time attributes. This will do the trick!

jacobcybulski

You are correct that "Replace Missing Values" is not appropriate for time series. Instead use "Replace Missing Values (series)", in which you'll have to define which of your attributes is an index (time) and which of the attributes need to be interpolated (using the previous and the following data point in time). You will also have an option to decide how the replacement happens. Jacob

anildas

Thanks Jacob, I have tried with REplace Missing Values(series), but I am not getting the required result,
attached a sample process for your understanding .. since i am very new to RM. I dont have much familiarity with different process available in RM for different purpose,
01-jan20 
03-jan20

required result set
01-jan-20
02-jan20
03-jan-20

Thanks
Anil

anildas

Thanks Jacob for providing me the hint/ option.
Instead of "Equalize Numerical Indices ", I have used "Equalize Time Stamps" operator, that exactly produce the required result.