reduce size of large network portal 06-26-2015d1047

2015-06-26

reduce size of large network portal 06-26-2015d1047
network size reduction
network reduction

Strategies for reducing the size of a large network while still approximating and maintaining many of the properties (mostly the property of the shape of the degree distribution) of the original network
-randomly remove edges
-remove an edge one by one to keep the graph as close to the original degree distribution according to a Pearson correlation coefficient as possible
-remove an edge one by one to keep the power law exponent as close to the original degree distribution value as possible.

Can the size of a scale free network be reduced in a manner that maintains many properties by simply removing all nodes with a degree less than a certain value? 06-26-2015d1044
The answer seems to be no.  I tried this with a large scale-free network and just removed all nodes with a degree less than 30, and the new distribution did not follow a power law. 
-file here:
--http://azim58.ngrok.io/files/ws-cnio/DR/2015/06%20June/06-09-2015d1021/successful%20creation%20of%20coexpression%20network/test%20reduce%20network%20size%20by%20removing%20nodes%20with%20degree%20less%20than%20k

---------
See also
-[cni_worked on network project to reduce network size by power law exponent 05-26-2015d1018]

azim58wiki: