Hi
The reference 18 in "Understanding the Modeling of Computer Network Delays using Neural Networks" points to the KDN site as the location of the code use:
http://knowledgedefinednetworking.org
I think that the section being referred to is labelled “Neural Nets:”, but I would like to confirm that please.
Many thanks
Nathan
Hi
In the demo code, for example in:
https://github.com/knowledgedefinednetworking/demo-routenet/blob/master/dem…
A form of normalisation is applied:
predictions = 0.54*preds + 0.37
This is consistent with the parse() function here:
https://github.com/knowledgedefinednetworking/demo-routenet/blob/master/cod…
if k == 'delay':
features[k] = (features[k] - 0.37) / 0.54
if k == 'traffic':
features[k] = (features[k] - 0.17) / 0.13
if k == 'link_capacity':
features[k] = (features[k] - 25.0) / 40.0
There is another case here:
https://github.com/knowledgedefinednetworking/network-modeling-GNN/blob/mas…
if k == 'delay':
features[k] = (tf.math.log(features[k]) + 1.78) / 0.93
if k == 'traffic':
features[k] = (features[k] - 0.28) / 0.15
if k == 'jitter':
features[k] = (features[k] - 1.5) / 1.5
if k == 'link_capacity':
features[k] = (features[k] - 27.0) / 14.86
if k == 'queue_sizes':
features[k] = (features[k] - 16.5) / 15.5
What is the basis for this normalisation and those specific constants please?
Why does the normalisation in routenet_with_link_cap and routenet_with_forwarding_nodes differ?
Many thanks
Nathan