search

ckchmod / rbn-matlab

Random Boolean Networks
0 stars 1 forks source link

Distance function #4

Closed darthbeeblebrox closed 7 years ago

darthbeeblebrox commented 7 years ago

Should take two different "boolCellGrid" classes and compare their final states in the "allStates" property, returning a Hamming distance.

Maybe this could just be a separate function.

Optional: could compare more than just the final states (e.g. if the grid is oscillating) and find some sort of "average distance"

ghost commented 7 years ago

I wrote a little function to calculate hamming distance of the final states and can look at comparing an average distance over more than just the final state.

darthbeeblebrox commented 7 years ago

Feel free to email it and I can add it in!

ghost commented 7 years ago

Hmmm just taking the average distance over multiple time steps doesn't necessarily give us new information if say two grids are oscillating between the same patterns but out of phase. Is there a way to find the optimal offset in time steps to minimize the distance?

darthbeeblebrox commented 7 years ago

Hmm, yeah, I'd thought about that problem but not what to do. It's also difficult if the cycles have different periods... I think there's something called the "Procrustes distance" for comparing different limit cycles, but I think for now we should only do steady states and not worry about cycles

ghost commented 7 years ago

Okay, we can file this away as a future improvement for now.

ckchmod commented 7 years ago

I agree with Charlie, if we have a network that we know arrives at steady state, we should pick that and do analysis on it. Just mention the caveat in the paper.

On Friday, December 2, 2016, darthbeeblebrox notifications@github.com wrote:

Hmm, yeah, I'd thought about that problem but not what to do. It's also difficult if the cycles have different periods... I think there's something called the "Procrustes distance" for comparing different limit cycles, but I think for now we should only do steady states and not worry about cycles

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/shinwookang/rbn-matlab/issues/4#issuecomment-264544267, or mute the thread https://github.com/notifications/unsubscribe-auth/AFItXxAUfkj1__NnVqLPZBXVA-96zF-3ks5rEHS6gaJpZM4LCAYV .

-- Best wishes, Chris Kang