Extracting and Scheduling Hotspots from Dynamic Spatial Environment for Recommendation

Abstract

Location-based services(LBSs) integrate one s geographic location with the application, bringing convenience to our daily lives and challenging industry and academia. LBSs have had substantial impacts on the way people travel and on multi-modal urban transport systems. These services include passenger transport, package delivery, and spatial crowdsourcing services, allowing the end-users to send requests from specific locations. Then the workers or the service providers move to that location to conduct the assigned tasks. The primary mechanism of all these service providers is task assignment, task scheduling, and resource management under different constraints like spatio-temporal constraints, capacity constraints, etc. newline newline newline newlineThe main challenges of location-based services are the optimal matching of resources (supply) and customers (demand), resulting in maximum resource utilization. Spatial recommendation systems play a significant role in achieving this goal of maximizing resource utilization. Using historical information, these spatial recommendation systems predict the upcoming demand in a region and suggest to the end-users or the workers the region with maximum utility. It enhances the user experience or the revenue of the system. In this thesis, we study the recommendation of Points of Interest(POIs) and hotspots to tourists and taxi drivers in four practical problems in spatial recommendation systems.