The spatiotemporal scan statistics proposed by Kulldorff have been applied to detect numerous disease clusters, and scan statistics based on heuristic algorithm optimization have also been utilized for disease cluster detection. The gravitational search algorithm (GSA) and the recent human mental search (HMS) possess superior performance in comparison with several heuristic algorithms, and neither algorithm has yet been applied in spatiotemporal scan statistics. However, the size of the spatiotemporal scanning window utilized in disease applications is constant in the time dimension, and it is difficult to detect changes in the size of an anomalous cluster over time. In this study, we proposed a dynamic cylinder with a variable radius as a spatiotemporal scanning window. In addition, we proposed an improved GSA based on mental search (MSGSA), and the MSGSA was utilized to optimize the dynamic scanning window to detect spatiotemporally anomalous clusters. The performance of the MSGSA was verified on 23 benchmark functions in comparison with the GSA and HMS. Simulated experiments based on the MSGSA and SaTScan showed that the MSGSA-optimized dynamic window yielded better performance based on the obtained accuracies and error rates. Finally, we utilized the MSGSA-optimized dynamic window and other methods to detect spatiotemporally anomalous clusters of hand-foot-and-mouth disease (HFMD) in China (2016) and Guangdong (2009), and the MSGSA-optimized dynamic window yielded better performance on both HFMD datasets. Moreover, the conclusions obtained with the MSGSA-optimized dynamic window were consistent with those of relevant researchers, indicating that the MSGSA possesses certain disease outbreak detection ability.

Big data applications are widely adopted to mine valuable information from a tremendous amount of industry data, which is commonly represented as a series of map-reduce operations. Among various map-reduce frameworks, Hadoop is most commonly adopted for data processing at large scale. Although Hadoop eases the development of highly scalable distributed big data applications, inefficient implementation due to poor coding practice and deep software abstractions can cause severe performance issues such as unreasonable slowdown, high response latency, and waste of computing resources, which can lead to unsatisfactory serving delay or significant maintenance cost. In this paper, we first categorize three common types of redundant patterns in big data applications. Then we propose a tool-assisted optimization workflow to detect the redundant patterns automatically, which profiles the application by sampling hardware performance monitoring units. Moreover, we present a profiling visualization method that can help to pinpoint the redundant codes. Based on these approaches, we optimize several big data applications by eliminating redundancies, yielding up to 14.8% performance improvement.