Multiple hypothesis testing in cognitive IoT sensor network

Recent Internet of Things (IoT) research aims to develop generic objects to learn, reason, and perceive their environment. Therefore, a new area has emerged known as cognitive IoT (CIoT). The cognitive Internet of Things integrates IoT with intelligence and behaves as well as humans through intelligent functionality. Several inferential tasks in CIoT require multiple hypothesis testing. The situation becomes cumbersome when the data is massive and heterogeneous. Thus, this research suggests a novel technique for multiple-hypothesis testing that uses a copula function to deal effectively with massive heterogeneous data. In addition, these data may contain missing or corrupted entries. Hence, it introduced probabilistic clustering, which reduces model inefficiency and takes control over the false discovery rate (FDR). Most of the variance from each cluster was extracted using kernel principal component analysis (KPCA) to reduce the processing burden at the fusion centre. Subsequently, it computes the p-value of each cluster's first principal component data and employs the Bonferroni method for multiple hypothesis testing. Finally, this research study evaluates the performance of the proposed algorithm on six-month environmental data, revealing that the proposed technique is efficient in terms of accuracy and computation time compared to other methods in the presence of massive heterogeneous data.