News for hackers. Stuff that matters.

A comment on Time Series Prediction – A short introduction for pragmatists. Parent comment

Re:Time Series Prediction – A short introduction for pragmat... by hprotagonist on Saturday November 09, 2019 @01:39PM (#21493339)
and not a single DS/ML person seems to have ever heard of it.
[ Parent | on HN ]
Re:Time Series Prediction – A short introduction for pragmat... by teej on Saturday November 09, 2019 @01:42PM (#21493357)
Because most folks in DS/ML are working on predicting likelihood to press a button. You don’t need Fourier transforms for that.
[ Parent | on HN ]
Re:Time Series Prediction – A short introduction for pragmat... by XuMiao on Saturday November 09, 2019 @02:56PM (#21493889)
CNN is essentially a FT with learnable kernels.
[ Parent | on HN ]
Re:Time Series Prediction – A short introduction for pragmat... by hprotagonist on Saturday November 09, 2019 @03:36PM (#21494181)
well, it’s certainly usually convolution. Hardly anyone knows why or how that connects to the FT though.
[ Parent | on HN ]
Re:Time Series Prediction – A short introduction for pragmat... by hprotagonist on Saturday November 09, 2019 @01:53PM (#21493443)
you do, depending on your feature preprocessing.
[ Parent | on HN ]
Re:Time Series Prediction – A short introduction for pragmat... by scottlocklin on Saturday November 09, 2019 @03:50PM (#21494267)
Most DS/ML people are useless chimps.

FWIIW I've used it all the time, along with wavelets, cepstrums, lifters and Hilbert transforms. For timeseries with nonlinearities and lots of data points, it's the way to go. 9/10 times I'd rather hire a EE with signal processing background than a statistician or data scientist for time series work.

[ Parent | on HN ]