Skip to main content
12 events
when toggle format what by license comment
Dec 26, 2021 at 3:50 comment added Khosraw Azizi All code for the model is here: github.com/Khosraw/… The function for the split is commented there for you to see. @Navin
Dec 23, 2021 at 18:42 comment added Navin 50MB is not large, it’s tiny. And there are two different “splits”: The first is your train/validation split (which might as well be 99%/1% + 100-fold cross validation since your dataset is so tiny). The other split is how many points you’re predicting on each eval (e.g. at inference the model sees 80 points and predicts the next 20 in the time series) which is pretty arbitrary and totally depends on your application. I think you’re confusing the two ratios. Post the code you used to split the data set and each time series in it.
Dec 11, 2021 at 4:36 vote accept Khosraw Azizi
Dec 8, 2021 at 0:31 comment added G__ Dont think the data type matters. You’ve already reported a substantial (28%) swing by changing the split. Above is a suggestion to help understand this.
Dec 6, 2021 at 13:51 history edited Khosraw Azizi CC BY-SA 4.0
added 35 characters in body
Dec 6, 2021 at 13:49 comment added Khosraw Azizi It's a time-series problem so not a lot will change.
Dec 5, 2021 at 17:43 comment added G__ Try repeating a few times w different 80/20 splits. See much variance? Maybe you got an unlucky split the first time…
Dec 5, 2021 at 13:01 answer added spectre timeline score: 4
Dec 5, 2021 at 12:37 history became hot network question
Dec 5, 2021 at 8:28 answer added eliangius timeline score: 5
S Dec 5, 2021 at 4:04 review First questions
Dec 5, 2021 at 5:14
S Dec 5, 2021 at 4:04 history asked Khosraw Azizi CC BY-SA 4.0