Spread dynamics and the confinement policies of COVID-19 exhibit different patterns for different countries. Numerous factors affect such patterns within each country. Examining these factors, and analyzing the confinement practices allow government authorities to implement effective policies in the future. In addition, they help the authorities to distribute healthcare resources optimally without overwhelming their systems. In this empirical study, we use a clustering-based approach, Hierarchical Cluster Analysis (HCA) on time-series data to capture the spread patterns at various countries. We particularly investigate the confinement policies adopted by different countries and their impact on the spread patterns of COVID-19. We limit our investigation to the early period of the pandemic, because many governments tried to respond rapidly and aggressively in the beginning. Moreover, these governments adopted diverse confinement policies based on trial-and-error in the beginning of the pandemic. We found that implementations of the same confinement policies may exhibit different results in different countries. Specifically, lockdowns become less effective in densely populated regions, because of the reluctance to comply with social distancing measures. Lack of testing, contact tracing, and social awareness in some countries forestall people from self-isolation and maintaining social distance. Large labor camps with unhealthy living conditions also aid in high community transmissions in countries depending on foreign labor. Distrust in government policies and fake news instigate the spread in both developed and under-developed countries. Large social gatherings play a vital role in causing rapid outbreaks almost everywhere. An early and rapid response at the early period of the pandemic is necessary to contain the spread, yet it is not always sufficient.

Health insurance is one of the most important services that people depend on for paying the bills related to hospital and clinical services. This dependency on health insurance lures some healthcare service providers to commit insurance frauds which has become a grave concern. The majority of healthcare fraud is committed by a very small number of untrustworthy providers. Yet, such fraudulent actions damage the reputation of the health service providers and cost the system billions of dollars. In this article, we specifically focus on the fraudulent claim identification problem and develop different solution schemes to identify the fraudulent cases in healthcare claims with minimal data. We present a solution to the fraudulent claim identification problem that translates diagnoses and procedure code's relations into Bipartite Graphs with Fuzzy Edges (BiGFuzzE). We also investigate the extension ofBiGFuzzEusing vector representations of clinical codes instead of non-negative matrix factorization (NMF). Our experimental evaluations demonstrate significant outcomes.

Patients depend on health insurance provided by the government systems, private systems, or both to utilize the high-priced healthcare expenses. This dependency on health insurance draws some healthcare service providers to commit insurance frauds. Although the number of such service providers is small, it is reported that the insurance providers lose billions of dollars every year due to frauds. In this article, we formulate the fraud detection problem over a minimal, definitive claim data consisting of medical diagnosis and procedure codes. We present a solution to the fraudulent claim detection problem using a novel representation learning approach, which translates diagnosis and procedure codes into Mixtures of Clinical Codes (MCC). We also investigate extensions of MCC using Long Short Term Memory networks and Robust Principal Component Analysis. Our experimental results demonstrate promising outcomes in identifying fraudulent records.

The simple question "Was this review helpful to you?" increases an estimated $2.7B revenue to Amazon.com annually 1. In this paper, we propose a solution to the problem of electronic product review accumulation using helpfulness prediction. The popularity of e-commerce and online retailers such as Amazon, eBay, Yelp, and TripAdvisor are largely relying on the presence of product reviews to attract more customers. The major issue for the user submitted reviews is to quantify and evaluate the actual effectiveness by combining all the reviews under a particular product. With the varying size of reviews for each product, it is quite cumbersome for the customers to get hold of the overall helpfulness.Therefore, we propose a feature extraction technique that can quantify and measure helpfulness for each product based on user submitted reviews.

Recent advances in Cyber-Physical Systems (CPSs) promote the Internet as the main communication technology for monitoring, controlling and managing the physical entities as well as exchanging information between the physical entities and human users. On the other hand, the Internet introduces a variety of vulnerabilities that may put the security and privacy of CPSs under risk. The consequences of cyber-attacks to CPSs might be catastrophic because they are usually part of human habitat. One of the most perilous threats in the Internet is the Denial of Service (DoS) attack and its variations such as Distributed DoS (DDoS). In this work-in-progress, we propose a novel probabilistic packet marking scheme to infer forward paths from an attacker to a victim site and delegate the defense to the upstream Internet Service Providers (ISPs). Our results show that the victim site can construct a forward path from the attacker after receiving 23 packets on the average.

In this work, we propose techniques to attain visibility into an arbitrary Internet subnetwork that is responsive to indirect probes but not to direct probes. By probing the network from a small number of selected vantage points, we are able to collect information about network-layer topology which would otherwise be hidden from measurement due to rate limiting practices, security mechanisms, and routing dynamics. We investigate the reasons for differing visibility, and the required number and placement strategies of vantage points needed to collect topology information at a low cost. We demonstrate substantial improvement in global visibility as probed by the TraceNET path measurement tool when leveraging only five vantage points selected according to route similarity.

Studying structural and functional characteristics of large scale graphs (or networks) has been a challenging task due to the related computational overhead. Hence, most studies consult to sampling to gather necessary information to estimate various features of these big networks. On the other hand, using a best effort approach to graph sampling within the constraints of an application domain may not always produce accurate estimates. In fact, the mismatch between the characteristics of interest and the utilized network sampling methodology may result in incorrect inferences about the studied characteristics of the underlying system. In this study we empirically investigate the sources of information loss in a sampling process; identify the fundamental factors that need to be carefully considered in a sampling design; and use several synthetic and real world graphs to elaborately demonstrate the mismatch between the sampling design and graph characteristics of interest.

Internet topology at the network layer consists of routers and subnets, i.e., point-to-point or multi-access connections. Network measurement studies have focused on router level maps and derived characteristics of routers such as mean degree, degree distribution, clustering coefficient and betweenness. Considering the fact that subnets are also important building blocks of the Internet topology, this paper introduces a complementary view of network topologies named subnet level maps. Subnet level network topology maps represent subnets as vertices and depict routers as links connecting the vertices/subnets. Additionally, we introduce a tool, called exploreNET, for subnet discovery. Although ExploreNET is based on the same principals as our recent work traceNET [21], it differs from traceNET in its utilization in various domains. Particularly, it allows us discover the underlying subnet level topology map of a network rather than the map dictated by routing dynamics. Finally, we present an evaluation of exploreNET by using it to discover and analyze various subnet characteristics including degree distribution, capacity distribution and utilization for six geographically disperse public Internet Service Providers (ISPs).

This paper presents a network layer Internet topology collection tool called tracenet. Compared to traceroute, tracenet can collect a more complete topology information on an end-to-end path. That is, while traceroute returns a list of IP addresses each representing a router on a path, tracenet attempts to return all the IP addresses assigned to the interfaces on each visited subnetwork on the path. Consequently, the collected information (1) includes more IP addresses belonging to the traced path; (2) represents "being on the same LAN" relationship among the collected IP addresses; and (3) annotates the discovered subnets with their observed subnet masks. Our experiments on Internet2, GEANT, and four major ISP networks demonstrate promising results on the utility of tracenet for future topology measurement studies.