Discovering Travel Spatiotemporal Pattern Based on Sequential Events Similarity

Algorithm 1

Travel sequence similarity algorithm (TSSA).

	Input: Travel sequence L1 and L2
	Output: Similarity of L1 and L2
	Initialization: Set score matrix M to 0
(1)	for i ⟵ 0 to |L1| do
(2)	M[i][0] ⟵ i  d2;
(3)	end for
(4)	for j ⟵ 0 to |L2| do
(5)	M[0][j] ⟵ j  d2;
(6)	end for
(7)	for i ⟵ 1 to |L1| do
(8)	for j ⟵ 1 to |L2| do
(9)	if Overlap (L1[i].p.type, L2[j].p.type) then, //Overlap (a, b) means the attraction type labels of POI a and POI b overlap
(10)	Spoi ⟵ PSA (L1[i].p, L2[j].p);
(11)	Stime ⟵ TSA (L1[i].t, L2[j].t);
(12)	sim ⟵ u  Spoi + (1 − u)  Stime;
(13)	M[i][j] ⟵ max(M[i − 1][j − 1] + sim, M[i − 1][j] + d2, M[i][j − 1] + d2);
(14)	else
(15)	M[i][j] ⟵ max(M[i − 1][j − 1] + d1, M[i − 1][j] + d2, M[i][j − 1] + d2);
(16)	end if
(17)	end for
(18)	end for
(19)	return M[|L1| − 1][|L2| − 1];