search

time-series-machine-learning / tsml-java

Java time series machine learning tools in a Weka compatible toolkit
GNU General Public License v3.0
158 stars 120 forks source link

timeseriesweka.classifiers #12

Closed TonyBagnall closed 5 years ago

TonyBagnall commented 5 years ago
  1. FastWWS: This is for Petijean DTW. It should be refactored to somewhere more sensible
  2. boss: these are variants for the IDA paper. I think they can go
  3. cote: Do we use these?
  4. ensembles. This seems a bit ad hoc
  5. randomboss. this should go in Matthews development branch (i.e. local). I think we have maybe two BOSS on the Repo: The original (BOSS) and our best effort so far (RandomBOSS)
    1. Classifiers: I suggest we oganise these in packagages dictionary, shapelet, frequencydomain, interval and timedomain
goastler commented 5 years ago

Is FastWWS related to EE? If so I can take a look at it, but could probably do with reading the paper on it also if you've got that to hand / can point me in the right direction

TonyBagnall commented 5 years ago

FastWWS is the Fast Window Search, which is a way of speeding up DTW window setting. This is the paper

TonyBagnall commented 5 years ago

https://www.francois-petitjean.com/Research/Petitjean2018-SDM-learn-warp-window.pdf

TonyBagnall commented 5 years ago

bloody newlines are annoying in this system :) I tested his version vs my fast version and his was 12x faster at a cost of about 50x memory, usually a reasonable trade off. Somewhere I documented the experiments .....