Gaussian copula network inference

This application enables the user to perform network inference on heterogeneous variables by assuming a Gaussian copula structure. Even if the data set is high dimensional, the computational cost is reduced by using a pairwise likelihood. The calculations are performed via the heterocop R package.
LOAD YOUR DATASET
The application can automatically detect your variable type (continuous or discrete), but we do not guarantee full reliability. You can check the box below to specify the number of variables of each type yourself. In this case, note that the columns of your .csv have to contain all continuous variables first.
DATASET
Which rows and columns do you want to display?
Columns
Rows
Download the selected dataset
The methodology for estimating the correlation matrix of the Gaussian copula is presented in the following preprint. Supplementary material is available on hal inrae .
Which rows and columns do you want to display?
Columns
Rows
Download the correlation matrix of the copula
Download the corrplot
The graph corresponding to the correlation matrix is shown below. An edge is present between two vertices if the absolute value of their correlation coefficient is greater than the threshold specified below (default 0.5).
The default colors correspond to the variable type (discrete or continuous), but more coloring options are available by ticking the box below.
Download the graph
The methodology for estimating the precision matrix of the Gaussian copula is presented in the following preprint. Supplementary material is available on hal inrae .
Which rows and columns do you want to display?
Columns
Rows
Download the conditional covariance matrix of the copula
Please tick the Optimal lambda selection box
Download the HBIC plot
The graph corresponding to the conditional covariance matrix is shown below. An edge is present between two vertices if their conditional covariance coefficient is non null. The default colors correspond to the variable type (discrete or continuous), but more coloring options are available by ticking the box below.
Download the graph