search

kperry2215 / unsupervised_anomaly_detection_time_series

This script pulls the gasoline price time series (from the EIA), and performs unsupervised time series anomaly detection using a variety of techniques. Techniques include SESD algorithm, One Class SVM, Isolation Forests, and low pass filter.
12 stars 2 forks source link

Comparisons #1

Open vgoklani opened 5 years ago

vgoklani commented 5 years ago

Hey there - I enjoy reading your blog posts, and was curious if you've compared any of the classical approaches to some of the deep learning techniques. That might make for a very interesting blog post!

kperry2215 commented 5 years ago

Vishal,

I’m glad you brought that up! The next post I’m working on is on automated model selection using H2O, and it will compare multiple ML and DL techniques. It may be valuable to add in a couple other posts though, such as forecasting using LSTM’s, and supervised anomaly detection in time series.

Thanks for the tip, and thanks for reading!

Kirsten

On Tue, Aug 27, 2019 at 8:08 AM Vishal Goklani notifications@github.com wrote:

Hey there - I enjoy reading your blog posts, and was curious if you've compared any of the classical approaches to some of the deep learning techniques. That might make for a very interesting blog post!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kperry2215/unsupervised_anomaly_detection_time_series/issues/1?email_source=notifications&email_token=AML4H7B46IGQ62STTSHAAVDQGUROHA5CNFSM4IQEIXW2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HHUXEQA, or mute the thread https://github.com/notifications/unsubscribe-auth/AML4H7DZNSWXZQ5GSKL66ADQGUROHANCNFSM4IQEIXWQ .