Taxi Fare Prediction System Using Key Feature Extraction in Artificial Intelligence

Abstract

Ridesharing is a service that uses travel information to match passengers, thus reducing the total demand of cars on road. However, the problems with the system are that it is expensive, and not suitable for high capacity and long distances. In this paper, we define the problems mentioned above methodologically. The proposed system uses information such as ride requests to achieve efficient taxi-request indexing and thus, improving the matching of taxi and passenger. Particularly, it marks indexes on taxis such as the geographical location and the source and destination, while selecting the most optimum route for the travel and satisfying both online, and offline ride bookings. Considerable amount of evaluation shows the accuracy of the system proposed, which is quick enough to process the requests in milliseconds. Unlike other services, it makes it easier to be deployed of shared and distributed infrastructure.