A Real-Time Taxicab Recommendation System Using Big Trajectories Data

Procedure 3

Distance Dispersion calculation and optimal recommendation.

(1) Input and abstract clustering results to new RDD.

(2) Using transformation, the carpool performance of a single potential destination of a real-time taxicab in Occupied

Status is quantified (Distance Dispersion) by user-defined functions.

(3) Using transformation, the carpool performance of a single real-time occupied taxicab is obtained

by aggregating all Distance Dispersion of potential destinations in .

(4) By the transformation, the descending order of the carpool performance (the ascending order of Distance

Dispersion) is processed.

(5) Using take(1) operation, we obtain and recommend the real-time taxicab in Occupied Status with the best

carpooling performance to the passenger .