Mapping Bus and Stream Travel Time Using Machine Learning Approaches

Algorithm 2

PseudoCode for implementation of SVM.

(1)	Model Parameters: C (Regularization/Penalty parameter), (width parameter/kernel coefficient), Epsilon ; Kernel-type: RBF;
(2)	Input: [bus travel time, time of the day, day of the week, no. of bus-stops];
(3)	Output: [stream travel time];
(4)	For every in the training dataset:(a)Transform the points into a kernel space using the kernel equations as defined,(b)Calculate the distance from the points to a hyperplane in the n-dimensional space such that .(c)Apply a penalty for points that do not satisfy condition,(d)Optimize/minimize this distance to find the ideal hyperplane.
(5)	For every in the test dataset:
	Find closest value to the corresponding hyperplane for given set of predictors, .