Efficient Processing of Uncertain Data Using Dezert-Smarandache Theory: A Case Study

摘要

Dezert-Smarandache Theory (DSmT) of plausible and paradoxical reasoning has excellent performance when the data contain uncertainty or conflicting. However, the methods developed in DSmT are in general very computationally expensive, thus they may not be directly applied to multiple data sources with high cardinality. In this paper, we explore the feasibility of using DSmT in practical applications through a case study. Specifically, we propose a DSm hybrid model with reduced number of classes and thus low computational cost to analyze temperature and humidity data received from multiple sensors to determine comfort zones in a smart building. Data from each sensor is considered as individual evidence that can be uncertain, imprecise and even conflicting. Several types of combination rules are applied to calculate the total mass function. Then the belief, plausibility and pignistic probability are deduced. They are used as metrics for decision making to determine comfort levels of the monitored environment. Both simulation and real data experiments demonstrate that the proposed method would make DSmT feasible for practical situation awareness applications.