ai-fast-track/blog

Time Series Classification Using Deep Learning - Part 1 | AI Fast Track Blog

Opened this issue · 4 comments

Time Series Classification Using Deep Learning - Part 1 | AI Fast Track Blog

A gentle introduction to time series classification using state of the art neural network architecture.

https://ai-fast-track.github.io/blog/timeseries/2020/05/21/time-series-using-deep-learning-part-1.html

Excellent article, Farid Hassainia, I particularly liked how you seamlessly integrated your package in fastai2! I've wanted to look into AI Time Series analysis for awhile, and your post provides a way to quickly get started in this field!

An awesome and quality blog post. I am very excited and pumped for the coming posts in the series. Something which are gonna really really helpful. Could you please tell, are there any regression datasets under URLs_TS? Does the same ts_learner with InceptionTime architecture handles the regression task and how can it be done with an example(4 lines of code)?
Thank you & Stay safe.

Thank you for the detailed introduction! "TSData.from_arff()" is an awesome function which can be used to read arff file. However, I am not sure how to generate a file like "NATOPS_TRAIN.arff". I know how to convert my data to the file like "NATOPSDimension1_Train.arff", which only contains one dimension. How can I combine different dimensions together? Thank you!

You might check out this video tutorial on how to create an arrf file. You can ignore the second part of the tutorial about WEKA:

Creating ARFF Files for Weka

I hope you will find it helpful!