doloMing / Encoding-decoding-and-causality-between-complex-networks

This is the toolbox introduced in the paper entitled as "Encoding, decoding, and causality between complex networks"

Encoding, decoding, and causality between complex networks

This is the toolbox introduced in the paper entitled as "Encoding, decoding, and causality between complex networks".

In general, this toolbox can be used to deal with the following questions: A. Represent a complex network by a Gaussian variable, where the infomation of network topology is contained in the covariance matrix of the Gaussian variable B. Analytically calculate the similarity between networks from the perspectives of encoding, decoding, and causal analyses C. Help solve network embedding, clustering, query, and classification

In the released MATLAB files, you can find key functions used in our research:

1. GRandomVariable.m: The function used to work out the graph Laplacian (the discrete Schr"{o}dinger operator), the Moore–Penrose pseudoinverse of L, the covariance matrix of Gaussian variable
2. NetworkApproximation.m: The function used to realize network approximation if two networks have different size
3. InfoDivergence.m: The function used to work out the information divergence between two networks
4. MutualInfo.m: The function used to work out the mutual information between two networks
5. FisherInfo.m: The function used to work out the Fisher information between two networks
6. GrangerCandTransferE.m: The function used to work out the Granger causality and transfer entropy between two networks

Apart from these functions, we also provide three random network models used in our research:

7. BAModel.m: The function used to generate a Barab'{a}si–Albert network
8. ErdosRenyi.m: The function used to generate a Erdos-Renyi network
9. WattsStrogatz.m: The function used to generate a Watts-Strogatz network Please note that these models are developed by previous studies cited in our paper.

There is also a supporting function used to generate multivariate Gaussian variables:

10. mvg.m: The function used to generate samples of a given multivariate Gaussian variable

To help users understand how to use our functions, we provide a simple instance in the released files:

11. Instance.m: The function used to run the simple instance

An equivalent implementation based on Python is also provided, which has the same functions as the MATLAB version and is much easier to use.