MOPAS and Busan City complete development of ‘AI-based passenger disembarkation information estimation analysis model’
Used for new construction and adjustment of public transportation routes... Laying the foundation for scientific transportation policy
An AI model has been developed that can predict up to 99% of the number of bus passengers and their points of departure.
Through this, it is expected that it will be highly useful in reorganizing public transportation routes in local governments that need measures to improve the efficiency of public transportation line operation.
The Ministry of the Interior and Safety and Busan Metropolitan City announced on the 21st that they have completed the development of an AI-based passenger disembarkation information estimation analysis model to support reasonable public transportation route reorganization.
The model developed this time was designed with the main functions of calculating transportation demand close to the actual through estimating passenger drop-off points and number of passengers, and finding potential demand for public transportation.
The AI analysis model estimates up to 3% of the number of people getting off at each route and stop through a three-step process.
In the first stage, AI learns passenger data with disembarkation information and predicts the disembarkation point of passengers without disembarkation information through a prediction algorithm (deep neural network, DNN).
If it is difficult to predict the drop-off point in the first step, the drop-off point is predicted in the second step using the residence estimation method (home-based analysis).
Step 3 is the passenger history tracking method. The stop at which most other passengers boarding at the same stop get off is assumed to be the drop-off point.
MOPAS utilized approximately 3 million pieces of public and private data, including transportation card usage history data, mobile carrier floating population data, and credit card usage data, to derive potential transportation demand.
This can be used to evaluate the rationality of existing operating routes and open late-night bus routes.
MOPAS expects that the model developed this time will be widely used in the scientific route reorganization process by local governments.
Local governments, which have had difficulty reorganizing routes to reflect actual transportation demand due to a lack of drop-off information, are expected to increase the convenience of residents' lives by establishing a foundation to reflect actual demand based on data.
It is expected that the government will also be able to use it to develop effective transportation policies based on the exact number of passengers for each public transportation route.
Kim Jun-hee, Director of MOPAS Public Data Bureau, said, “It is significant in that it laid the foundation for a scientific transportation policy by discovering through data analysis the volume of passengers that had been difficult to determine until now,” adding, “We will continue to enhance administrative capacity through data and improve the lives of the people in a practical way.” “We will work hard to change,” he said.
Inquiries: Ministry of the Interior and Safety Digital Government Office Integrated Data Analysis Center (044-205-2289), Busan Metropolitan City Digital Economy Innovation Office Big Data Statistics Division (051-888-2545)