Hi,
Anyone who has used Cytoscape might be familiar with the Network Analyzer tool in Cytoscape which is used to calculate the degree distribution of the network. I did the same, plot the degree distribution and fit a power law y = ax(squared b) where b is the slope. Cytoscape gives the R-squared and correlation values of the network. By my understanding, The correlation coefficient (R) is squared to obtain the coefficient of determination (R-squared). However, none of my networks have the R-squared value as a square of the correlation value. For example, this is the correlation and R-squared values of three of my networks.
R-squared Correlation
0.684 0.963
0.532 0.942
0.621 0.949
Am I getting this wrong and does correlation mean something different and not correlation coefficient (R) in Cytoscape?
•
written
by
The Last Word
If you're fitting a power law, the coefficient of determination is most likely calculated on log(data). This is probably mentioned somewhere in the docs.
EDIT: Found it mentioned in the docs here.
Hi Jean, just to clarify, the correlation refers to the correlation coefficient right? It just says correlation and I am confused as to if that stands for something else.
Yes this is the correlation coefficient.
Hello....It supports now three different protocol encryptions: 0.75 (very first SimpleModulus), Season 0 - 6(E1), Season 6E3 (v1.04d gmo). - Load/Save of your recorded sessions - Simple packet sender - Improved network stack with reduced memory allocations and increased speed (uses byte Spans and System.IO.Pipelines) - Minor fixes/improvements
Log in to answer this question.
