Trajectory Clustering in an Intersection by GDTW

Algorithm 4

(ADPC_make_decision_graph).

	ADPC_make_decision_graph
	Input: D: normalized asymmetric generalized DTW
	r: cutoff-distance
	Return: rho: local density list
	delta: minimal distance list.
(1)	rho, delta = list[len(D)], list[len(D)]
(2)	for i = 0…len(D)-1
(3)	neighbors = D[i][D[I]<=r]4
(4)	rho[i] = gaussan_kernel(neighbors)
(5)	sortedidx = argsort(rho)
(6)	for i = 0…len(D)-1
(7)	r i = sortedidx[i]
(8)	if i = = len(D)-1
(9)	delta[ri] = D[:, ri].max()
(10)	else:
(11)	idx = sortedidx[i + 1:end]
(12)	delta[ri] = D[:, ri][idx].min()
(13)	return rho, delta