Regression model for the prediction of perceived wait time
In order to expand upon these results, a regression model for the prediction of perceived wait time based on the measured
wait time, presence of mobile real-time information, and other bus and rider factors was developed. Possible variables included
an indicator variable for gender, an indicator variable for college-age, a categorical variable for frequency of rider
use as well as indicator variables for the categories of rider frequency, a discrete variable for the rider aggravation level, a
continuous variable of the approximate distance the rider was about to travel, environmental categorical variables for
weather, time of day categorical variables, a discrete variable for bus frequency/headway, and a continuous variable for
the percent of late buses on the route.
After testing these variables for their significance, the final model was determined to be: