High-Speed Data-Driven Methodology for Real-Time Traffic Flow Predictions: Practical Applications of ITS

Algorithm 1

Pseudocode for the KNN-S2S prediction algorithm.

Given the without , and values, value:

(where )

Step 1: Build -NN

If (the previous ) then -value = 1

(where is the present running time; is a run-time limit)

(1) Initialize the list of the -objects shown as (5).

(2) For each and in the past database

(where ; )

(2-1) Calculate between and by eq. (4)

(2-2) If then

(where )

(2-2-1) Delete and from the -object list

(where is associated with , )

(2-2-2) Update and into the -object list

(2-2-3) Search in the updated -object list

(3) Sort the select -objects ascending based on

Step 2: Decide -NN and estimate

If then

If is available then -value = 0

(1) Add to

(2) Recalculate all in the -object list as follows

(where )

(3) For each and in the - object list

(where )

If then

(where )

(3-1) Delete and from the -object list

(where is associated with , )

(3-2) Update and into the - object list

(3-3) Search in the updated -object list

Else -value = 1

(4) Estimate by eq. (6)