Hierarchical clustering for multiple nominal data streams with evolving behaviour

Over the decade, a number of attempts have been made towards data stream clustering, but most of the works fall under clustering by example approach. There are a number of applications where clustering by variable approach is required which involves clustering of multiple data streams as opposed to clustering data examples in a data stream. Furthermore, a few works have been presented for clustering multiple data streams and these are applicable to numeric data streams only. Hence, this research gap has motivated current research work. In the present work, a hierarchical clustering technique has been proposed to cluster multiple data streams where data are nominal. To address the concept changes in the data streams splitting and merging of the clusters in the hierarchical structure are performed. The decision to split or merge is based on the entropy measure, representing the cluster’s degree of disparity. The performance of the proposed technique has been analysed and compared to Agglomerative Nesting clustering technique on synthetic as well as a real-world dataset in terms of Dunn Index, Modified Hubert $$\varGamma $$ Γ statistic, Cophenetic Correlation Coefficient, and Purity. The proposed technique outperforms Agglomerative Nesting clustering technique for concept evolving data streams. Furthermore, the effect of concept evolution on clustering structure and average entropy has been visualised for detailed analysis and understanding.

Edge Computing for Secured IoT Analytics on the Cloud

This chapter provides an insight into building healthcare applications that are deployed in the cloud storage using edge computing and IoT data analytics approaches. Data is collected from environments both within or external to the hospital. The devices that are connected enable the healthcare providers to monitor patients at large distances, manage chronic disease, and manage medication dosages. The data from these devices can be added to clinical research to gain an insight into the participant's experiences. Artificial intelligence techniques like machine learning or deep learning can be employed at the edge of the networks for IoT analytics of multiple data streams in online mode. The industrial edge computing is growing rapidly from 7% in 2019 to being expected to reach approximately 16% by 2025. The total market for intelligent industrial edge computing that includes hardware, software, services has reached $11.6B in 2019 and is expected to increase to $30.8B by 2025.

Characterizing super-spreading events and age-specific infectivity of COVID-19 transmission in Georgia, USA

As the current COVID-19 pandemic continues to impact countries around the globe, refining our understanding of its transmission dynamics and the effectiveness of interventions is imperative. In particular, it is essential to obtain a firmer grasp on the effect of social distancing, potential individual-level heterogeneities in transmission such as age-specific infectivity, and impact of super-spreading. To this end, it is important to exploit multiple data streams that are becoming abundantly available during the pandemic. In this paper, we formulate an individual-level spatiotemporal mechanistic framework to statistically integrate case data with geo-location data and aggregate mobility data, enabling a more granular understanding of the transmission dynamics of COVID-19. We analyze reported cases from surveillance data, between March and early May 2020, in five (urban and rural) counties in the State of Georgia USA. We estimate natural history parameters of COVID-19 and infer unobserved quantities including infection times and transmission paths using Bayesian data-augmentation techniques. First, our results show that the overall median reproductive number was 2.88 (with 95% C.I. [1.85, 4.9]) before the state-wide shelter-in-place order issued in early April, and the effective reproductive number was reduced to below 1 about two weeks by the order. Super-spreading appears to be widespread across space and time, and it may have a particularly important role in driving the outbreak in rural area and an increasing importance towards later stages of outbreaks in both urban and rural settings. Overall, about 2% of cases may have directly infected 20% of all infections. We estimate that the infected children and younger adults (<60 years old) may be 2.38 [1.30, 3.51] times more transmissible than infected elderly (>=60), and the former may be the main driver of super-spreading. Through the synthesis of multiple data streams using our transmission modelling framework, our results enforce and improve our understanding of the natural history and transmission dynamics of COVID-19. More importantly, we reveal the roles of age-specific infectivity and characterize systematic variations and associated risk factors of super-spreading. These have important implications for the planning of relaxing social distancing and, more generally, designing optimal control measures.