dc.description.abstract | A bike-sharing system is a service in which a fleet of bicycles is made available to the public
on a short-term basis through self-served docking stations. These stations are limited
in capacity and are often depleted or saturated with bikes due to sudden spikes in demand.
These spikes are hard to avoid and are both detrimental to the user experience and the
effectiveness of the system. Machine learning methods have been used to forecast demand
spikes at station level in similar systems successfully and would likely be a valuable tool
in proactively counteracting the effect of demand spikes in the Oslo bike-sharing system.
The goal of this thesis is to evaluate common machine learning methods for demand prediction
modeling at individual bike-sharing stations in Oslo.
To accomplish this, four machine learning methods which have successfully predicted
station-level demand in similar systems are evaluated through a set of experiments. The
methods are based on a random forest, gradient boosting tree and recurrent neural networks
with either long short-term memory- or gated recurrent unit units.
Based on the experimental results, the recurrent neural network with the long short-term
memory unit is deemed to be the most suitable for the Oslo bike-sharing system, both
due to performance and future potential. However, all methods achieved good performance
and made accurate predictions.
The results pave the road for developing a full-scale prediction system in the Oslo bike-sharing
system, by highlighting the most promising prediction method. | |