So I have this data from the previous device release (old model). Date / # subscribers
Old Model Data
Month 1: 100
Month 2: 200
Month 3: 300
Month 4: 350
Month 5: 350
Month 6: 350
I want to predict how the new model will fare based on the old model curve. (e.g. next 4 months)
New Model Data
Month 1: 50
Month 2: 150
Any suggestions appreciated.
And to be clear, I have fit a model to the "old" data and attempted to apply it to the new data, but it doesn't seem to work very well. And yes, I am using the predict command.
The actual "old" data has about 18 data points. New has 3 data points.
Your sample size is too small to get a semi-accurate result for the New Model Data.. Try expanding your sample size
Have you looked at the predict
command yet?
... | timechart count(Month) as count | predict count
http://docs.splunk.com/Documentation/Splunk/6.5.3/SearchReference/Predict