物联网搜索中主客观融合的动态信誉评估

与传统的互联网搜索相比，物联网搜索更加强调数据质量。为了提高数据质量，物联网搜索中心需要依数据质量为提供者支付相应报酬以激励其持续提供符合质量需求的数据。这就使如何评估数据提供者的数据质量及其信誉成为物联网搜索中的一个基础问题。为此，引入动态信誉机制来综合评估数据提供者的可信性，为数据选择、收集提供依据。首先，提出了主观意愿及客观质量评价计算方法，并设计了交互行为贴现及信誉衰减机制；然后，提出了融合主客观要素的动态信誉计算模型，并基于信号传递机制进行博弈分析，以获取交互次数、贴现率、支付价格及数据成本之间约束关系。实验证明该模型能够较好地反映数据质量变化情况，为数据选择提供依据，并满足物联网搜索的实时性及动态性需求。

Fusing subjective and objective factors: a dynamic approach to evaluating reputation for IoT search

Compared with the traditional Internet search, IoT search data center needs a higher data quality. In order to effectively motivate data owners to continuously provide higher quality data, the center needs pay the corresponding rewards to them according to the quality of data (QoD). As a result, QoD and the trustworthy evaluation become a basic problem for the development of IoT search. To address this problem and support data selection of IoT center, a dynamic reputation model was proposed to comprehensively assess the reputation of data owners. In detail, first, an approach to assessing the subjective and objective quality was proposed and mechanism of interactive discount and reputation attenuation was designed. Then, fusing subjective and objective factors, a novel dynamic reputation evaluation scheme was presented. Last, in order to get constraint relation of interaction times, discount, payment price, and data cost, an economic analysis based on signaling game was conducted. Experiment results show that the proposed approach can reflect dynamic change of QoD effectively, and provide the basis for data selection. It also can be adapted to real-time and dynamic of IoT search.