Machine-learning to inspire Singapore metro buildout

"Singapore needs an efficient transport system to support people's activities given the existing and planned infrastructure," said researcher Christopher Monterola.
By Brooks Hays   |   Jan. 25, 2017 at 12:29 PM
share with facebook
share with twitter

Jan. 25 (UPI) -- Researchers are trying to distill smart transit philosophy into a machine-learning algorithm. Scientists hope their smart transit model will reveal a recipe for a smarter city, organized in way that relieves the congestion common on the mass transit systems of major cities.

"Singapore needs an efficient transport system to support people's activities given the existing and planned infrastructure," project leader Christopher Monterola, a researcher at the Agency for Science, Technology and Research's Institute of High Performance Computing, explained in a news release. "To guide planners, we needed a model that could predict ridership under the regional centers plan."

Like many cities, Singapore consists of a large central downtown, or an inner central business district, surrounded by less dense residential and industrial zones. With so many commuting in and out of the central business district at rush hour, the setup promotes congestion.

Planning officials are working to promote less centralized urban density -- regional centers spread throughout the city state.

To predict how these efforts and other land use trends will affect metro ridership and transportation patterns, researchers have turned to machine-learning.

Scientists supplied their algorithm with both ridership and land-use distribution data. Researchers plotted the paths of more than 20 million bus and subway journeys over the course of week. They combined ridership patterns with information on the concentration of lands used for business, industry, residence and outdoor recreation.

The researchers experimented with three machine-learning models to see which best predicted the relationship between land-use and ridership.

"We found that a decision tree model performed best, with good accuracy, computational efficiency and an easy-to-follow user display," Monterola explained. "Results indicated that an increase in amenities of up to 55 per cent across the city would increase ridership. Beyond this point, ridership begins to decline; this is logical because if amenities are available locally, people walk instead."

Dense concentrations of amenities were the best predictors of mass transit use. Researchers hope their findings -- detailed in the journal Land Use Policy -- will help city officials expand and augment the mass transit system to better meet and anticipate the needs of Singapore's riders.

Related UPI Stories
Trending Stories