Big Data Applications - Big Data: Related Technologies, Challenges and Future Prospects - page 66

analysis in SNS is also known as social media analysis. Social media include
text, multimedia, positioning, and comments. Nearly all research topics related to
structural analysis, text analysis, and multimedia analysis may be interpreted as
social media analysis, but social media analysis is confronted with unprecedented
challenges. First, massive and continually growing social media data should be
automatically analyzed within a reasonable time. Second, social media data contains
much noise, e.g., blogosphere contains a large number of spam blogs, and so does
trivial Tweets in Twitter. Third, SNS are dynamic networks, which are frequently
and quickly changed and updated.
Since social media is close to SNS, social media analysis is inevitably influenced
by SNS analysis. SNS analysis refers to the text analysis of SNS context and
characteristics of social and network structures, as well as multimedia analysis. The
existing research on social media analysis is still in its infancy. The applications
of SNS text analysis include transfer learning in keyword search, classification,
clustering, and heterogeneous networks. Keyword search tries to synchronously use
contents and link behaviors for search [ 47 ]. The motivation for such applications
is that text files containing similar keywords are generally connected to each
other [ 48 ]. During classification, assuming all nodes of the SNS are provided with
labels, the nodes added with labels are classified. During clustering, researchers
aim to determine node sets with similar contents and accordingly group them [ 49 ].
Considering that SNS contains massive information of different interlinked objects,
e.g., articles, labels, images, and videos, transfer learning in heterogeneous networks
aims to transfer knowledge information among different links [ 50 ].
Multimedia datasets in SNS is organized in a structured form, which brings
rich information, e.g., semantic ontology, social interaction, community media,
geographical maps, and multimedia opinions. Structural multimedia analysis in SNS
is also called multimedia information networks. The link structure of multimedia
information networks is mainly a logic structure, which are of vital importance to
the multimedia in multimedia networks. The logic connection structures in multi-
media information networks can be classified into four types: semantic ontology,
community media, individual photo albums, and geographical positions [ 36 ].

6.2.6 Mobile Traffic Analysis

With the rapid growth of mobile computing, mobile terminals and applications in
the world are growing rapidly. By April 2013, Android Apps has provided more
than 650,000 applications, covering nearly all categories. By the end of 2012, the
monthly mobile data flow has reached 885 PB [ 51 ]. The massive data and abundant
applications exploit a broad research field for mobile analysis but also bring about
a few challenges. As a whole, mobile data has unique characteristics, e.g., mobile
sensing, moving flexibility, noise, and a large amount of redundancy. Recently, new
research on mobile analysis has been started in different fields. Because of the far