Re: ForceAtlas publication
Posted: 15 Feb 2011 23:07
By the way,
I finally got the principle of the "distributed attraction" feature. The possibility to twist the mechanism of the layout, to make it depend on the topology of the network, opens many possibilities. For example, does it mean that, following the same principle, it would be possible to scale the attractive or repulsive forces according to attributes of the nodes?
Say, I have an attribute ranging from 1 to 10: nodes with an attribute value of 10 would be 10 times more attractive than nodes with att value of 1.
Again, following this logic, node's dynamic attributes (say, their date of start and date of end) could be a factor playing a role in the determination of attractive and repulsive forces - great possibilites I guess. Like, younger nodes in the network would have a lesser attraction force, which would scale up to the mean with time passing.
All that makes me cringe because I can't code in Java! Would love to try all those ideas!! ))
I finally got the principle of the "distributed attraction" feature. The possibility to twist the mechanism of the layout, to make it depend on the topology of the network, opens many possibilities. For example, does it mean that, following the same principle, it would be possible to scale the attractive or repulsive forces according to attributes of the nodes?
Say, I have an attribute ranging from 1 to 10: nodes with an attribute value of 10 would be 10 times more attractive than nodes with att value of 1.
Again, following this logic, node's dynamic attributes (say, their date of start and date of end) could be a factor playing a role in the determination of attractive and repulsive forces - great possibilites I guess. Like, younger nodes in the network would have a lesser attraction force, which would scale up to the mean with time passing.
All that makes me cringe because I can't code in Java! Would love to try all those ideas!! ))