Multiview topological data analysis for crowdsourced service supply-demand gap prediction

Abstract

The widespread of smart devices equipped with high sensing capabilities opened the door to the new paradigm of mobile crowdsourcing. This paradigm relies on the crowd contribution and participation to collect data and relevant information. The abstraction of mobile crowdsourcing as a service has become easier and more seamless thanks to the availability, low-cost and fast access to cloud services. In this context, it is important to satisfy a request for a crowdsourced service, at a given time and place, as soon as possible. Nevertheless, maintaining a balance between the supply and demand of crowdsourced services in a geographic area is challenging given the mobility of both service requesters and providers. Motivated by this requirement, we propose a forecasting approach to infer the supply demand gap of crowdsourced services in a given geographic area. Instead of relying on raw data for prediction, we devise a technique to generate predictors from the raw gap data using topological data analysis to exploit the topological and underlying geometric structures. Our forecasting strategy is conducted in a multiview fashion, that is, we devise the historical time horizon into immediate, near and distant time. Then, using topological analysis, we derive three key features: topological similarity, Betti numbers and the Distance To Measure value (DTM). These features, along with additional context information including weather and temperature, are used to infer the supply-demand gap value using state-of-art prediction approach. Our experiments show that the proposed multiview topological analysis is effective for supply-demand prediction with both clean and noisy data. 2020 IEEE.