How to predict the waiting time of a person in a minibus queue?
Author: Tom
Abstract
This research aims to develop a prediction model that can predict the waiting time of a person in a minibus queue.
Research Context
The waiting time of a person depends on the eta(estimated time of arrival) of the minibus and the number of minibus that the person needs to wait for. For example, if the person can get on the next minibus (i.e. the number of minibus he/she needs to wait for is 1), then the waiting time is equal to the eta of the next minibus. If the person needs to wait for 2 minibusses, the waiting time is equal to the eta of the 2nd next minibus. Therefore, we can predict “the number of minibus that the person needs to wait for” to get the waiting time of that person.
Researching Method- Data Collection
We collected the data on the “number of empty seats of the next minibus” of the minibus route ‘11’(To Choi Hung) in HKUST for 3 days. With the data, we can know if a person in the queue can get on the next minibus. Then we can generate a dataset with “index of the person in the queue” as the input parameter and “can the person get on the next minibus” as the output.
Prediction Model
Using naïve Bayes, we can predict the number of minibuses that a person needs to wait for. Integrate the data with the eta from the government API (https://data.etagmb.gov.hk/eta/route-stop/2004791/1/7), we can get the predicted time of the person.
Results
