A study on Time Series using sequence to sequence model in TensorFlow.
This repo follows the original post from:
weiminwang.blog/2017/09/29/multivariate-time-series-forecast-using-seq2seq-in-tensorflow/
You can also look at the ipython notebook for code details.
Contents -
- Univariate time series¶
- Multivariate time series
- Seq2seq for outliers/extreme events
- A case study - Beijing pollution data
Dataset credits - https://archive.ics.uci.edu/ml/datasets/Beijing+PM2.5+Data