Multivariate-Time-Series-forecast-using-seq2seq-in-TensorFlow

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 -

1. Univariate time series¶
2. Multivariate time series
3. Seq2seq for outliers/extreme events
4. A case study - Beijing pollution data

Dataset credits - https://archive.ics.uci.edu/ml/datasets/Beijing+PM2.5+Data