Multivariate Travel Time Forecasting in a Traffic Network Using Fuzzy Cognitive Mapping

Traffic forecasting is a crucial component in the implementation of efficient traffic management and control. While various works have been made in this area, most of the current works focus on traffic flow forecasting. Although traffic flow is an important feature in traffic modeling, other variables such as travel time, are also useful in developing robust traffic models. Travel time forecasting is a relatively underexplored area in the literature, especially when additional features are considered in generating the forecasts. This paper uses the fuzzy cognitive mapping approach to develop a multivariate travel time forecasting model. The construction of the FCM uses the pseudoinverse learning algorithm. A case study was performed in a road network in Cebu, Philippines with 20 selected routes. The results of the study showed that the FCM model outperforms the benchmark models in terms of directionality and magnitude. The paper significantly contributes to the literature in three ways. Firstly, it is one of the few works that investigate multivariate travel time forecasting. Secondly, it pioneers the application of FCM to multivariate traffic forecasting. Finally, it is one of the few works offering lenses in the Philippines. The findings in this study would be beneficial for stakeholders, such as traffic practitioners, policymakers, and urban planners in the development of relevant initiatives.