IEEE Systems, Man and Cybernetics Magazine - April 2020 - 39

and requires extensive training and expertise. They used
natural-language processing to classify pieces of text as
either argumentative or nonargumentative and clustered
the text fragments in the hope that these clusters would
contain similar arguments. The VAT image was used to
assess clustering tendency in the data set; although it
showed some small clusters, it was inconclusive in suggesting major cluster structure in the data. This prompting
the authors to conclude that the analysis cannot go far
without an extensive pretagged corpus.
Building Linguistic Summaries From the Sensor Data
As information technology advances, more and more data
are created, stored, and analyzed. However, this vast
mountain of data is beyond human cognitive capabilities
and comprehension skills. Hence, methods to summarize
data and analyze these summaries are becoming increasingly important. Several approaches for linguistic summarization have been proposed in the literature, which
generates linguistic summaries from sensor data so that
people can read and take appropriate action. For instance,
summaries of sensor data on older residents in independent-living facilities, including n
- ighttime motion activity
and restlessness while lying in bed, provide indications of
potential abnormal conditions [130], [131].
As the number of sensors grows, so does the complexity
and size of the set of linguistic descriptions. Hence, it is
necessary to perform some automated analysis to condense this information. A set of papers by Wilbik et al.
[132]-[134] develop an approach to generating linguistic
prototypes from a group of time blocks that represent a
normal condition from a care environment for older individuals. Then, the set of summaries for new time blocks is
compared to the prototypes to flag anomalous conditions,
thereby reducing the burden on the human. Wilbik et al.
developed novel distance measures between linguistic
summaries and use VAT/iVAT to assess and validate cluster
structure in the linguistic summaries, which allows for the
creation of linguistic prototypes from clusters of summaries over some temporal range. Anomalies are detected as
observations that are considerably different from the linguistic prototypes in a moving temporal window.
Web User Data Analysis
The World Wide Web generates a humongous amount of
data in the form of weblogs, user activity, browsing preference, social media, user-generated content, and so on. Web
analytics is the measurement, collection, analysis, and
reporting of web data for purposes of understanding and
optimizing web usage. However, web analytics is not just a
process for measuring web traffic; it can be used as a tool
for business and market research and to assess and
improve the effectiveness of a website. Various researchers have used some members of the VAT family of algorithms to analyze web-generated data to solve different
problems. We turn to this application next.

Modeling User Behavior
Clustering web sessions to identify visitors' choices while
browsing web pages is an important problem in web mining.
The sequence of pages viewed by a user in a particular time
frame (i.e., the session) captures his/her interest in a specific topic. The clustering of these sessions can be used to provide customized services. Chakraborty and Bandyopadhyay
[135] and Sisodia et al. [136] explored the use of clustering
web sessions to provide customized services to users with
similar interests. The VAT algorithm was used to visualize
the clustering tendency of these data, which was later fed as
an input to the actual clustering algorithm.
Sun et al. [137] worked on the problem of detecting
threats to the security, privacy, and integrity of computer
networks and infrastructure from insiders-those who
have (or had) authorized access to an organization's network, system, or data and intentionally exceeded or misused that access in a manner that negatively affected the
confidentiality, integrity, or availability of the organization's
information or information systems. The authors modeled
system users' behavior and developed fast and efficient
techniques to predict and detect insider threats. A necessary step for this task is to understand legitimate resourceusage access patterns, which can help identify abnormal or
suspicious user behavior. Cluster analysis based on visual
assessment using the VAT algorithm enabled them to detect
communities of users based on the projects they access and
the networks they use. Based on normal behavior patterns
characterized by different clusters, abnormal behavior,
detected by observing a deviation from normal user behavior, provides a pathway to developing an insider-threat
detection system.
An important component of web design is user privacy
because people are disclosing more and more personal information on online platforms. However, there is a problematic
gap between existing online privacy controls and actual user
disclosure behavior, which motivates researchers to focus
on the design and development of intelligent privacy controls. Intelligent controls decrease the burden of privacy
decision making and generate user-tailored privacy suggestions. The first necessary step is to analyze user privacy preferences. Kaskina [138] used VAT to assess clustering
tendency in this context and then applied a fuzzy clustering
approach to a real-world data set collected from a political
platform. The fuzzy membership degree values were used to
calculate more precise personalized privacy suggestions.
Fraud Detection
On social media networks, automated social agents (i.e.,
bots) are increasingly becoming a problem. Fraud in bot
messaging, email spam, opinion spam, and so on, is a
major threat to the credibility of web-based services and
applications. Spam or unwanted email is one of the potential issues of Internet security, and classifying user emails
correctly from penetration of spam is an important issue
for antispam researchers.
Ap ri l 2020

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE

IEEE Systems, Man and Cybernetics Magazine - April 2020

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - April 2020