Good afternoon. Noticed a bug in implementation, probably.
I suppose you should update a weight of a neighbour n by moving it to the observed point x.
So that way L81 should be placed inside for loop and be changed to store update for node n, not s_1
Am I wrong?
|
update_w_s_n = e_n * (np.subtract(observation, self.network.node[s_1]['vector'])) |
|
for neighbor in self.network.neighbors(s_1): |
|
self.network.node[neighbor]['vector'] = np.add(self.network.node[neighbor]['vector'], update_w_s_n) |
Good afternoon. Noticed a bug in implementation, probably.
I suppose you should update a weight of a neighbour n by moving it to the observed point x.
So that way L81 should be placed inside for loop and be changed to store update for node n, not s_1
Am I wrong?
GrowingNeuralGas/gng.py
Lines 81 to 83 in abd12c6