标题：Approach of measuring PoC of context using limited self-feedback in context-aware systems
作者：Ji, Mingyang ;Xu, Hongji ;Wang, Leitao ;Dang, Juan ;Xu, Zhengzheng ;Fang, Haiteng
作者机构：[Ji, Mingyang ;Xu, Hongji ;Wang, Leitao ;Dang, Juan ;Xu, Zhengzheng ;Fang, Haiteng ] School of Information Science and Engineering, Shandong Universit 更多
来源：IET Wireless Sensor Systems
摘要：In context-aware systems (CASs), due to various reasons such as faulty sensors, the context information received from a context provider does not always represent accurately the reality, and incorrect contexts often lead to conflicts in multi-source conditions. The need of accurate data fusion results in research on the parameter - probability of correctness (PoC), which denotes the probability that a piece of information is correct. This study focuses on precise evaluation of PoC and utilises this parameter to accomplish more accurate data fusion ultimately. The authors propose a novel approach based on limited self-feedback by taking the output of data fusion process as criterion to evaluate PoC automatically and dynamically. Supervisory mechanism is used to ensure the credibility of self-feedback data. The updated PoC will affect the accuracy of the following rounds of data fusion based on Dempster-Shafer (DS) theory. Experiments testify from multiple perspectives that compared with the traditional user feedback-based method, this self-assessment-based approach can achieve more precise evaluation of PoC with less human interaction and effectively improve the accuracy of data fusion through PoC-based DS theory. Thus, the information or services provided by CASs will be more accurate and efficient. © The Institution of Engineering and Technology 2016.