Relieving the Congestion around a School via Automatic Vehicle Identification Technology-Based Traffic Measures

Algorithm 1

Agglomerative hierarchical clustering.

	Input: a set of feature vectors for individual behaviour portrait, the number of elements , with log-likelihood distance measure.
	Output: license plate numbers of two types of vehicles .
(1)
(2)	Classify each element in into a class
(3)	Calculate the similarity matrix for the classes in
(4)	for to do
(5)	Merge the two closest samples as a new class;
(6)	The other existing classes in ;
(7)	Update the similarity matrix;
(8)	Until two types of group left.
(9)	end for
(10)	return the clustering license plates group