/GP_TSE

Primary LanguageJupyter NotebookMIT LicenseMIT

GP_TSE

The code and data for the paper:

image

Requirements

  • GPflow >= 2.0.0
  • tensorflow >= 2.0.0
  • tensorflow_probability
  • Other packages: numpy, scipy, matplotlib
  • Some baseline models, adaptive smoothing interpolation (ASM) and Spatiotemporal Hankel Low-Rank Tensor Completion (STH-LRTC) are implemented in MATLAB.

Data

The data used in the paper is available at data, the data in mat and npy formats have the same values.

Getting Started

  • See demo.ipynb for a quick start.
  • The experiments in the paper can be reproduced by running the scripts in experiments. Note there could be some errors about the path of the data, please modify the path in the scripts to the correct path of the data.