An Entrenching Recommender System for Sustainable Transportation: A Transit to Terminus

Introduction

Selecting the most efficient route for travel is a common occurrence in the daily lives of many people. There are numerous tools available for this use. Systems called route planners assist users in choosing the most efficient path to take to get to their destinations(Jiang et al., 2019; Li & She, 2017; Nayak & Panda, 2018; C. D. Wang et al., 2019). Multi-modal route planners incorporate various modes of transportation into one trip, including walking, car- and bicycle sharing, and private and public transportation. For instance, a commuter can use park and rides to transition from a private vehicle to public transportation and then use a bike-sharing Programme to get to work after exiting the subway(Yun et al., 2018). Modern route planning tools like Google Maps and Apple Maps typically ignore user preferences or the wisdom of the crowd when recommending multi-modal routes, instead only providing the quickest or shortest routes between two randomly chosen places. Route design should not be viewed as a classic shortest path problem, hence understanding this is crucial(Panda et al., 2020). Route planners could instead make the route recommendations more user-friendly. By incorporating user feedback into the route generating process, it may be possible to find routes that will leave users feeling the most satisfied in a certain situation(Hartatik et al., 2018; Herzog et al., 2017; Huang et al., 2019; H. Liu et al., 2019). Locals who frequently use public transportation are more knowledgeable about the best routes to take to get where they're going. This is especially useful when it's busy out, with congested streets and crowded public transportation. Having access to such information enables commuters to choose less congested routes and modes of transportation(Herzog et al., 2017; Huang et al., 2019). A possible approach to addressing the shortcomings of modern route planners is the integration of recommender systems (RSs). The information overload issue can be effectively solved by using RSs, which are software tools and processes that select objects like movies or restaurants that are most likely to be of interest to the user(Nakamura et al., 2014). Collaborative filtering (CF), which makes recommendations to users based on their shared characteristics with other users, is one of the most used RSs approaches(Bajaj et al., 2016; Y. Liu et al., 2021; Nakamura et al., 2014; Nawara & Kashef, 2020; J. Wang et al., 2018). CF is domain-free since it bases its suggestions exclusively on how frequently people provide the same feedback on products and engage with the system. In this chapter, we present, a unique mobile routing system-based recommender system (RS) for individualized, multimodal journeys. We demonstrate how applying CF to the wisdom of the crowd can improve the quality of route recommendations. Furthermore, we describe how to expand this strategy by adding a knowledge-based component to get around the pure CF recommendations with some future accessible imminent.