VisibilityGraphMotifs_Encoder

Encode the motifs of the VisibilityGraph into a given motif size.

Encoders in language Python and Matlab are provided.

#Python #Matlab #Visibility Graph

Chinese:

可视图的编码器。可以将所给的时间序列翻译成可视图。提供python和matlab两种语言。

文件的注释为英文。关于可视图和编码的算法请参考本文的最末的参考文献。

如果有任何问题，可以提出issues。欢迎使用本代码进行学术研究，别忘了引用本仓库的链接，随手给个star也行哦！:)

---

Our paper entitled "Natural visibility encoding for time series and its application in stock trend prediction" has been published 😃.

If you are interested in our work and having difficulties in accessing the article, please contact me.

---

Contents

This repository contains the following files:

NVGcode.m

• In Matlab.
• Encodes the time series (sequence) into the natural visibility graph series. (For any provided motif size.)
• Using the maximum slope idea, a little bit faster than the direct processing. Still O(n^2).

HVGcode.m

• In Matlab.
• Encodes the time series (sequence) into the horizontal visibility graph series. (For any provided motif size.)
• Using the maximum slope idea, a little bit faster than the direct processing. Still O(n^2).

NVGcode_DC.m

• In Matlab.
• Encodes the time series (sequence) into the natural visibility graph series. (For any provided motif size.)
• Using the "divide and concur" idea, O(NlogN). Details refer to Fast transformation from time series to visibility graphs.

NVGcode_DC.py

• In python.
• Encodes the time series (sequence) into the natural visibility graph series. (For any provided motif size.)
• Using the "divide and concur" idea, O(NlogN). Details refer to Fast transformation from time series to visibility graphs.

Reference

If you use the codes, Please star this repository and cite the URL of this repository. Please~ :)

You could also cite our paper using:

@article{HUANG2021107478,
title = {Natural visibility encoding for time series and its application in stock trend prediction},
journal = {Knowledge-Based Systems},
volume = {232},
pages = {107478},
year = {2021},
issn = {0950-7051},
doi = {https://doi.org/10.1016/j.knosys.2021.107478},
author = {Yusheng Huang and Xiaoyan Mao and Yong Deng},
}

If you have any problems, please do not hesitate to contact me.

The reference of the visibility graph are provided as follows:

• From time series to complex networks: The visibility graph The core paper of the Natural Visibility Graph.
• Horizontal visibility graphs: Exact results for random time series The core paper of the Horizontal Visibility Graph.
• Fast transformation from time series to visibility graphs An O(NlogN) method, using the "divide and concur" philosophy.