social web mining
TRANSCRIPT
-
8/7/2019 Social Web Mining
1/8
c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1623
j o u r n a l h o m e p a g e : w w w . i n t l . e l s e v i e r h e a l t h . c o m / j o u r n a l s / c m p b
Social Web mining and exploitation for serious applications:
Technosocial Predictive Analytics and related technologies
for public health, environmental and national security
surveillance
Maged N. Kamel Boulos a,, Antonio P. Sanfilippo b, Courtney D. Corley b, Steve Wheeler c
a Faculty of Health, University of Plymouth, Drake Circus, Plymouth, Devon PL4 8AA, UKb Techno-Social Predictive Analytics Initiative, Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352, USAc Faculty of Education, University of Plymouth, Drake Circus, Plymouth, Devon PL4 8AA, UK
a r t i c l e i n f o
Article history:
Received 3 December 2009
Received in revised form
10 February 2010
Accepted 12 February 2010
Keywords:
Social Web
Web 2.0
Disease surveillance
Technosocial Predictive Analytics
a b s t r a c t
This paper explores Technosocial Predictive Analytics (TPA) and related methods for Web
data mining where users posts andqueries aregarneredfrom SocialWeb(Web 2.0) tools
such as blogs, micro-blogging and social networking sites to form coherent representations
of real-time health events. The paper includes a brief introduction to commonly used Social
Web tools such as mashups and aggregators, and maps their exponentialgrowth as an open
architecture of participation for the masses and an emerging way to gain insight about
peoples collective health status of whole populations. Several health related tool examples
are described and demonstrated as practical means through which health professionalsmight create clear location specific pictures of epidemiological data such as flu outbreaks.
2010 Elsevier Ireland Ltd. All rights reserved.
1. Background
The rapid development and growing popularity of the
Social Web, also commonly known as Web 2.0, and its
applications, coupled with the increasing availability of
mobile Internet-enabled devices (such as netbooks and AppleiPhone), has ensured that the Internet is now a truly
integral part of everyday life. The Social Web is a partic-
ularly dynamic medium that captures the pulse of the
society in real time. Blogs, micro-blogs such as Twitter
(http://twitter.com/), and social networking sites such as
Corresponding author at: Room 05, 3 Portland Villas, Faculty of Health, University of Plymouth, Drake Circus, Plymouth, Devon PL4 8AA,UK. Tel.: +44 0 1752 586530; fax: +44 0 7053 487881.
E-mail addresses: [email protected] (M.N. Kamel Boulos), [email protected] (A.P. Sanfilippo),[email protected] (C.D. Corley), [email protected] (S. Wheeler).
Facebook (http://www.facebook.com/), enable people to pub-
lish their personal stories, opinions, product reviews and
a great deal more in real time. They can also share geo-
tagged alerts and reports about their current location and
allow others to follow their whereabouts, all in real time,
using location-aware social services such as Microsoft Vine
(http://www.vine.net/). Such applications have greater power
when coupled to Social Web mashup services and aggre-
gators. Such tools weave data from disparate sources into a
new, compound data source or service [1,2]. A useful exam-
ple of the latter is HealthMap, the global disease alert map
(http://www.healthmap.org/en).
0169-2607/$ see front matter 2010 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.cmpb.2010.02.007
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://twitter.com/mailto:[email protected]:[email protected]:[email protected]:[email protected]://www.facebook.com/http://www.vine.net/http://www.healthmap.org/enhttp://dx.doi.org/10.1016/j.cmpb.2010.02.007http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://www.healthmap.org/enhttp://www.vine.net/http://www.facebook.com/mailto:[email protected]:[email protected]:[email protected]:[email protected]://twitter.com/http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
2/8
c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3 17
Fig. 1 How To Use Who Is Sick? (Screenshot from http://whoissick.org/sickness/).
The dynamic nature and continuously updated features of
the Social Web make it a fertile environment for intelligencegathering in a variety of disciplines, enabling users to tap into
the wisdom of the crowds. Further, the Social Web is useful
for gleaning the collective impression of the masses regard-
ing current matters, events and products [3]. In fact, it is not
uncommonthese days for example,to see shoppers(including
those who do not make their purchases through the Internet)
consulting online user reviews and ratings on sites such as
Amazon.com about products they are planning to buy before
making their purchase decisions.
This paper will briefly review a number of emerging
technologiesand tools, includingTechnosocialPredictiveAna-
lytics pioneered at Pacific Northwest National Laboratory,
USA (http://predictiveanalytics.pnl.gov/ ), that can be used toexploit these inherent Social Web features in real or near-real
time and harness them for public health, environmental and
national security surveillance purposes.
2. A brief introduction to TechnosocialPredictive Analytics and some relatedtechnologies and tool examples
2.1. Who Is Sick: the power of user-generated maps
Who Is Sick (http://whoissick.org/sickness/ , Fig. 1) was
launched in 2006 as a location-aware service for tracking andmonitoring sickness, with the aim of providing current and
local sickness information by the public to the public. People
can anonymously post information about their own illness
and use a familiar Google Maps interface to search and fil-
ter illness by symptoms, sex, age, and location. The maps act
as a crude syndromic surveillance tool in real time. A person
feeling unwell can instantly check what sicknesses are circu-
lating within their own locale. Health-conscious individuals
conduct similar scans and take anynecessary preventive mea-
sures in preparation. People traveling to another area of the
country or abroad can use the service to learn if there are any
sicknesses going around that they need to be aware of. One
should be warned however, that information quality is always
dependent upon the accuracy of user-reported details and on
sufficient numbers of people (who are actually sick) using thetool to report their symptoms in a timely manner.
2.2. We Feel Fine: differentiating between informative
and affective text on the Social Web
Created in 2006, We Feel Fine (http://www.wefeelfine.org/)
is a Web-based applet and API (Application Programming
Interfacehttp://wefeelfine.org/api.html ) that can be used to
assess the fluctuating effects of real world events and factors
such as the Stock Index or the weather on peoples feelings
and emotions, and ascertain a crude idea about the general
mood of populations at different dates and locations around
the world [4].At the heart of We Feel Fine is a data crawler
that automatically scours the Web every 10min, har-
vesting data in the form of human feelings from a
large number of blogs and social pages such as those
hosted by MySpace (http://www.myspace.com/) and Blogger
(http://www.blogger.com/). We Feel Fine scans blog posts for
occurrences of the phrases I feel and I am feeling, and
once any of these is found, the system goes backward to the
beginning of the sentence and forward to the end of the sen-
tence, and then saves the full sentence in a database. Saved
sentences are next scanned against a manually compiled list
of about 5000 pre-identified feelings (adjectives and some
adverbs). If a valid feeling is found, the corresponding sen-tence is considered to represent one person who feels that
way. We Feel Fine also attempts to extract the username of
theposts author from the URLof his/her blog post (e.g., steve-
wheeler in http://steve-wheeler.blogspot.com/ ) and uses it
to automatically traverse the blog hosting site to locate that
user profile page. From the profile page, it is often possible to
extract the age, gender, country, state, and city of residence
of the blog owner. We Feel Fine can also retrieve the local
weather conditions for the blog owners city at the time the
post was written. This process results in about 15,000to 20,000
saved feelings (with associated data) per day. For display pur-
poses, the top 200 feelings were manually assigned colours
that loosely correspond to the tone of the feeling, e.g., happy
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://whoissick.org/sickness/http://predictiveanalytics.pnl.gov/http://whoissick.org/sickness/http://www.wefeelfine.org/http://wefeelfine.org/api.htmlhttp://www.myspace.com/http://www.blogger.com/http://steve-wheeler.blogspot.com/http://steve-wheeler.blogspot.com/http://www.blogger.com/http://www.myspace.com/http://wefeelfine.org/api.htmlhttp://www.wefeelfine.org/http://whoissick.org/sickness/http://predictiveanalytics.pnl.gov/http://whoissick.org/sickness/http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
3/8
18 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3
Fig. 2 Screenshot of the available filtering options (feeling, gender, age, weather, location [country city], date) in the We
Feel Fine applet (the applet can be accessed by clicking on Open We Feel Fine link athttp://www.wefeelfine.org/).
positive feelings are graphically represented using bright yel-low and sad negative feelings are depicted in dark blue. The
applet also offers a number of filtering options for querying
the underlying database (Fig. 2).
While We Feel Fine functions as intended with its
pre-coordinated search strategy and list of feelings, it is
rather limited and cannot fully capture all possible affective
expressions and differentiate between them and informative
sentences in Social Web text. One can easily appreciate how
complex this task is by searching for some keywords: such
as flu at BingTweets (http://bingtweets.com/ ) and observing
the nonstop tweets in the resulting real-time Twitter feed
pane: there will almost always be people tweeting about
their own flu sickness (e.g., back to bed to try and sleep offthis flu. My sinuses feel numb!counts as one flu case) and
others just sending informative tweets about flu (sharing
a news item, some related research or information, or ask-
ing a question, e.g., How business travelers can avoid swine
flu: http://bit.ly/zBOuYdoes not count as a flu case). The
challenge here is to devise an algorithm that can reliably dif-
ferentiate between the two types of tweetsautomaticallyand
in real time. Ni et al. [5] describe one solution using a machine
learning method for classifying informative and affective blog
posts.
We Feel Fine also attempts to perform some geographical
classification of the feelings it continually gathers from the
blogosphere, but again it does so in a rather limited fashion.
The limitations extend beyond looking at users profiles forlocation details (which are not always available), because there
may also be issues to individual posts. A post by, for example
a user from the USA, might relate to that users experience at
a totally different location to the one they are normally resid-
ing in, e.g., the Bloggers holidays in Italy. Fortunately, more
sophisticated technologies exist today, such as MetaCarta
(http://www.metacarta.com/), which can be applied to sift
through Social Web text and extract much more compre-
hensive geographical references, while intelligently handling
any geographical name ambiguities or inconsistencies in the
scanned text [6].
2.3. Google Flu Trends: an example of infodemiologyand infoveillance in action
Eysenbach [7,8] defines infodemiology as the science of
distribution and determinants of information in an elec-
tronic medium, specifically the Internet, or in a population,
with the ultimate aim to inform public health and pub-
lic policy. He further goes on to define a related concept,
infoveillance, as a similar science whose primary aim is
surveillance. Infodemiology and infoveillance rely on auto-
mated and continuous analysis of unstructured, free text
information available on the Internet. Thisincludes analysisof
search engine queries (or what Eysenbach calls the demand
side), as well as of what is being published on Web sites, blogs,
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://www.wefeelfine.org/http://bingtweets.com/http://bit.ly/zBOuYhttp://www.metacarta.com/http://www.metacarta.com/http://bit.ly/zBOuYhttp://bingtweets.com/http://www.wefeelfine.org/http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
4/8
c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3 19
Twitter, etc. (or the supply side, to use again Eysenbachs
terminology).
Building on the same infodemiology and infoveillance
conceptsdescribed by Eysenbach, researchers at Google found
that certain search terms are good indicators of flu activity
[9]. Google Flu Trends (http://www.google.org/flutrends/) uses
aggregated Google search data to estimate flu activity up to
two weeks quicker than traditional systems. Such an earlydetection of disease activity (in close to real time), when fol-
lowed by an appropriate rapid response, has the potential of
reducing the impact of both seasonaland pandemic influenza.
Trends (continuously updated) are currently available from
Google for Australia, New Zealand, the United States (Fig. 3),
Mexico and sixteen more countries. A related project, Infovigil
(http://www.infovigil.com/ ), has recently been launched by
Eysenbach and colleagues at the Centre for Global eHealth
Innovation, Toronto, Canada, and is introduced in [8]. (Seealso
the Technosocial Predictive Analytics approach to monitoring
of seasonal influenza later in this paper.)
2.4. Emergency and public health virtual situationrooms using 3D serious gaming technology: getting the
big picture in real time
Conventional situation rooms are routinely used to oversee
public health emergencies and disaster management opera-
tions in realtime. Nowadays, large amounts of emergency data
are increasingly coming from a wide range of sources in real
or near-real time and need to be cross-linked and visualized
where they spatially belong on maps of the affected regions.
Also, emergencies are usually managed by multi-professional
teams who, not uncommonly, are distributed in multiple geo-
graphic locations. Because of these reasons, we have started
to see physical situation rooms gradually being replaced (orcombined) with virtual situation rooms that use onlinecollab-
orative (and mostly 2Dtwo-dimensional) platforms such as
Depiction (http://www.depiction.com/), in addition to conven-
tional Web conferencing. These platforms, although usable
and helpful, leave much to be desired, as they are lacking the
third dimension, which is needed to create a proper percep-
tion of the emergency space, as well as a sense of co-presence
of other virtual team members [1,10,11].
To overcome this limitation, Kamel Boulos [11,12] pro-
posed the development of novel emergency and public health
virtual situation rooms in a suitable 3D (three-dimensional)
virtual world (sometimes also called 3D serious gaming
platform), where animated 3D graphical self-representations(avatars) of experts and professionals can collaborate together
and discuss incident data in real time in a simulated 3D
space representing the physical location where the emer-
gency/public health incident of interest is unfolding and
reflecting all changes taking place there (again in real time).
The real-time link between the virtual world and the physical
world incident would be two-way and multimodal involving
geo-tagged physical/environmental sensor data feeds, citizen-
contributed data (as found in Microsoft Vine and Who Is
Sick), data gleaned via automatic analysis of Social Web
content, textual and 3D spatialized audio/voice exchanges
between virtual team members, video feeds, 3D simulations
and animations, various Web mashups, and shared desk-
top applications, among other possibilities. Various what-if
scenarios can also be explored and tested in the virtual
environment, making the proposed rooms much suited for
real-time emergency and disaster management, e.g., for man-
aging an influenza pandemic and planning and coordinating
actions at global, regional and local levels.
To achieve the above vision, the proposed collabo-
rative and interactive situation room platform wouldmarry virtual globes or 3D mirror worlds (such as
Google EarthTMhttp://earth.google.com/) and 3D virtual
worlds (such as Second Lifehttp://secondlife.com/ and
OpenSimhttp://opensimulator.org/ ) [1,11,12], and comple-
ment and tightly integrate them with other key technologies,
e.g., real time, geo-tagged RSSReally Simple Syndication
data feeds and geo-mashups (using Web services such as
Yahoo! Pipeshttp://pipes.yahoo.com/ ). The platform would
weave data and services in real time from different sources
into a new rich datascape that better reflects the current
situation (the big picture) in novel visual ways that are easier
to understand and manage. The platform has to be secure,
enabling multiple distributed persons to see each other,visualize relevant data together in unique ways, conduct
3D simulation and training scenarios, and collaborate in
real time, each according to their assigned role and access
privileges. True stereoscopic vision can be added, if needed,
using readily available technologies such as NVIDIA 3D Vision
(http://www.nvidia.com/object/3D Vision Overview.html) for
more realistic 3D visualization, with a bettersense of 3D depth
and relief. It is noteworthy that it is now possible to stream a
3D virtual world to a suitable mobile phone (see, for example,
http://www.youtube.com/watch?v=XwRnjbkljnc ) or other
Internet-enabled, small form factor mobile devices, making
this vision end-user device and platform-independent and
thus suitable for those members of the emergency operationsteam who are on the move.
2.5. Technosocial Predictive Analytics: a systems
science approach to proactive anticipatory reasoning
Events occur daily that challenge the health, security and
sustainable growth of our society, and often find us unpre-
pared for the catastrophic outcomes. These events involve
the interaction of complex processes such as climate change,
emerging infectious diseases, energy reliability, terrorism,
nuclear proliferation, natural and man-made disasters, and
geopolitical, social and economic vulnerabilities. If we are to
prevent the adversities and leverage the opportunities thatemerge from these events, integrated anticipatory reasoning
has to become an everyday activity. Technosocial Predictive
Analytics (TPA [13]) supports anticipatory reasoning through a
multidisciplinary approach to strategic analysis and response
that enables a concerted decision-making effort by analysts
and policymakers.
There is increased awareness among subject-matter
experts, analysts, and decision makers that a combined
understanding of interacting physical and human factors is
essential in addressing strategic decision making proactively.
For example, multilevel studies that consider a broad range
of biological, family, community, socio-cultural, environmen-
tal, policy, and macro-level economic factors are necessary to
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://www.google.org/flutrends/http://www.infovigil.com/http://www.depiction.com/http://earth.google.com/http://secondlife.com/http://opensimulator.org/http://pipes.yahoo.com/http://www.nvidia.com/object/3D_Vision_Overview.htmlhttp://www.youtube.com/watch%3Fv=XwRnjbkljnchttp://www.youtube.com/watch%3Fv=XwRnjbkljnchttp://www.nvidia.com/object/3D_Vision_Overview.htmlhttp://pipes.yahoo.com/http://opensimulator.org/http://secondlife.com/http://earth.google.com/http://www.depiction.com/http://www.infovigil.com/http://www.google.org/flutrends/http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
5/8
20 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3
Fig. 3 Google Flu Trends for the United States (screenshot ofhttp://www.google.org/flutrends/intl/en us/). The site was
displaying a data current through: October 04, 2009 on the day this screenshot was taken (on October 05, 2009). Users canalso view flu activity by individual state.
prevent and mitigate the emergence of health threats such
as childhood obesity [14] and drug addiction [15]. Integrated
understanding of the infrastructural, social and ideational
context in which contentious social movements operate is
an essential analytical step in framing the emergence of
violent behaviour [16]. Combined understanding of anthro-
pogenic effects (e.g., chemical waste) and natural processes
(e.g., solar variation) is needed to predict the impact of global
warming [17]. Insights on human cognitive and emotional
biases improve our understanding of economic decisions andtheir effect on market prices, returns, and the allocation
of resources [18]. TPA provides theoretical and operational
advancements of these insights.
The human brain provides a basic framework for mem-
ory and prediction in which decision making emerges as a
process of pattern storage, recognition and projection rooted
in our experience of the world and driven by perception
and creativity [19]. Qualities such as the ability to focus on
what is perceived to be most important in the moment
and the capacity to make quick decisions by insight and
intuition make human judgment uniquely effective [20,21].
However, the same qualities can also be responsible for fal-
lacious reasoning when judgment lacks sufficient knowledge
and expertise [22], and is biased by factors such as group-
think [23,24], limitations in memory and focus attention
[25,26], positive framing [27] and increased confidence in
extreme judgments and highly correlated observables [28].
TPA addresses these gaps by using knowledge management
andmodeling to informhumanjudgment andanalytical gam-
ing to neutralize biased reasoning. Morespecifically, TPA helps
analysts and policymakers provide better proactive analysis
and response by enabling naturalistic decision making while
countering adverse influences on human judgment through acombined set of capabilities that (a) provide multidisciplinary
knowledge reach-back to inform analysis and response dur-
ing decision making; (b) supplement the expertise of the
analyst and policymaker with simulated scenarios generated
by computational models that integrate human and physi-
cal factors, and (c) engage analysts and policymakers within
a gaming environment that stimulates creative critical rea-
soning through visual analysis and collaborative/competitive
work, as described in Fig. 4a.
TPA offers a collaborative visual analytic environment
with user-friendly access to predictive models in the areas
of security, energy and the environment that are informed
by knowledge management processes, as shown in Fig. 4b.
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://www.google.org/flutrends/intl/en_us/http://www.google.org/flutrends/intl/en_us/http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
6/8
c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3 21
Fig. 4 Top diagram (a) describes TPAs capability components. Bottom diagram (b) illustrates TPAs operational
environment.
Such an environment promotes model integration through
innovative algorithms that enable the combination of diverse
modeling paradigms such as Bayesian networks and system
dynamics. Integrated modeling is informed by a Knowl-
edge Encapsulation Framework that facilitates the distillation
and vetting of relevant knowledge from heterogeneous
data streams, including social media. The TPA collaborative
working environment enables analysts and policymakers to
stress-test the validity of their analysis and policy plans using
techniques such as role-playing and serious gaming. Gaming
data are stored over time and analyzed to make inferences
about strategic decision making in context from social inter-
action during game-play.
TPA provides an ideal systems science approach for pub-
lic health informatics with reference to challenges such as
emerging infectious diseases. Current epidemic models char-
acterize effects of transportation, demographic traits, schools,
healthcare access, and host-to-hostdynamics. TPA can extend
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
7/8
22 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3
Fig. 5 CDC ILINet vs. normalized flu-keywords: containing ILI-based blog post frequency per week.
these existing modeling paradigms through the integration
of socio-cultural models to provide a more accurate char-acterization of how public health responses reflect factors
such as social order, individual and community-based health
knowledge, and behaviour change due to public health com-
munications in mass media and emerging Web technologies.
One specific application of the TPA approach to public
health informatics is the monitoring of seasonal influenza
through Web and social media. As described earlier in this
paper, many Internet users formulate search queries and post
entries in blogs and discussion forums using terms related to
influenza-like-illness (ILI) to identify sites and otherusers that
offer medical diagnosisand advice [9,29]. There is a clear sense
in which an increase/decrease in the number of ILI-based
searches and posts in blogs and discussion forums reflects ahigher/lower public concern for the spreading of influenzaand
can therefore be used to monitor seasonal influenza. Such an
insight is corroborated by observed correlations of ILI-based
blog posts and surveillance reports from sentinel health-
care providers. For example, Fig. 5 plots ILI-based blog posts
(source: http://www.spinn3r.com) with outpatient ILI Surveil-
lance Network (ILINet) by the US Centers for Disease Control
and Prevention on a weekly basis for the period from Octo-
ber 05, 2008 to January 25, 2009. The evaluation of Pearsons
correlation coefficient for the two data series yields a strong
(r = 0.545) and statistically significant (95% confidence) correla-
tion value. Such results and analyses can greatly improve the
quality of existing models of seasonal influenza.
3. Conclusions
Technosocial Predictive Analytics is still in its infancy, but
nevertheless provides health professionals with a powerful
method of exploiting vast, continuous streams of user-
generated content on a rapidly proliferating worldwide social
network. As the technology develops it will provide more
sophisticated and reliable tools for public health use. It must
be emphasized however, that such techniques are only appli-
cable where large proportions of a society regularly use Social
Web services. There are still many people and sometimes
entirecommunities whodo notuse theWeb on a regular basis.
For such communities there may be no easy way to gain gen-eral health data, and it is inadvisable to extrapolate findings
from other communities in an attempt to predict their health
state. However, as Social Web use increases exponentially due
to technology and connectivity becoming more widely avail-
able andaffordable, this situation will rapidly change. There is
therefore a need to monitor the continuously shifting demo-
graphic characteristics of a variety of Social Web services that
populations are using. Each social network service and its
prevailing user characteristics such as age, gender and preva-
lent user geographic locations should be considered as these
data can help in better interpreting any intelligence or results
derived from these services.
r e f e r e n c e s
[1] M.N. Kamel Boulos, M. Scotch, K.-H. Cheung, D. Burden, WebGIS in practice VI: a demo playlist of geo-mashups forpublic health neogeographers, International Journal ofHealth Geographics 7 (2008) 38, Available at: http://www.ij-healthgeographics.com/content/pdf/1476-072X-7-38.pdf(last accessed 2.12.09).
[2] S. Wheeler, M.N. Kamel Boulos, Mashing, burning, mixingand the destructive creativity of Web 2.0: applications formedical education, RECIISElectronic Journal ofCommunication, Information and Innovation in Health 1 (1)
(2007) 2733, Available at: http://www.revista.cict.fiocruz.br/index.php/reciis/article/view/49/51 (last accessed 2.12.09).
[3] M.N. Kamel Boulos, S. Wheeler, The emerging Web 2.0 socialsoftware: an enabling suite of sociable technologies inhealth and healthcare education, Health Information andLibraries Journal 24 (March (1)) (2007) 223.
[4] R. Mittal, S. Nikkila, Analyzing the effect of real world eventson feelings using we feel fine, Available at:http://www.public.asu.edu/rmittal2/docs/MittalNikkilaProjectReport.doc (last accessed 2.12.09).
[5] X. Ni, G.-R. Xue, X. Ling, Y. Yu, Q. Yang, Exploring in theweblog space by detecting informative and affective articles,in: Proceedings of the 16th International Conference onWorld Wide Web, Banff, Alberta, Canada (Session: IndustrialPractice & Experience), 812 May, 2007, pp. 281290,
Available at: http://www2007.org/papers/paper225.pdfor
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://www.spinn3r.com/http://www.ij-healthgeographics.com/content/pdf/1476-072X-7-38.pdfhttp://www.revista.cict.fiocruz.br/index.php/reciis/article/view/49/51http://www.revista.cict.fiocruz.br/index.php/reciis/article/view/49/51http://www.public.asu.edu/~rmittal2/docs/MittalNikkila_ProjectReport.dochttp://www.public.asu.edu/~rmittal2/docs/MittalNikkila_ProjectReport.dochttp://www.public.asu.edu/~rmittal2/docs/MittalNikkila_ProjectReport.dochttp://www.public.asu.edu/~rmittal2/docs/MittalNikkila_ProjectReport.dochttp://www2007.org/papers/paper225.pdfhttp://www2007.org/papers/paper225.pdfhttp://www.public.asu.edu/~rmittal2/docs/MittalNikkila_ProjectReport.dochttp://www.public.asu.edu/~rmittal2/docs/MittalNikkila_ProjectReport.dochttp://www.revista.cict.fiocruz.br/index.php/reciis/article/view/49/51http://www.revista.cict.fiocruz.br/index.php/reciis/article/view/49/51http://www.ij-healthgeographics.com/content/pdf/1476-072X-7-38.pdfhttp://www.spinn3r.com/http://dx.doi.org/10.1016/j.cmpb.2010.02.007 -
8/7/2019 Social Web Mining
8/8
c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 0 0 ( 2 0 1 0 ) 1 6 2 3 23
http://doi.acm.org/10.1145/1242572.1242611(last accessed2.12.09).
[6] M.N. Kamel Boulos, On geography and medical journalology:a study of the geographical distribution of articles publishedin a leading medical informatics journal between 1999 and2004, International Journal of Health Geographics 4 (March1) (2005) 7.
[7] G. Eysenbach, Infodemiology: tracking flu-related searches
on the web for syndromic surveillance, AMIA AnnualSymposium Proceedings (2006) 244248.
[8] G. Eysenbach, Infodemiology and infoveillance: frameworkfor an emerging set of public health informatics methods toanalyze search, communication and publication behavior onthe Internet, Journal of Medicine Internet Research 11(March (1)) (2009) e11.
[9] J. Ginsberg, M.H. Mohebbi, R.S. Patel, L. Brammer, M.S.Smolinski, L. Brilliant, Detecting influenza epidemics usingsearch engine query data, Nature 457 (February (7232)) (2009)10121014.
[10] G.F. Welch, D.H. Sonnenwald, H. Fuchs, B. Cairns, K.Mayer-Patel, H.M. Sderholm, R. Yang, A. State, H. Towles, A.Ilie, M.K. Ampalam, S. Krishnan, V. Noel, M. Noland, J.E.Manning, 3D medical collaboration technology to enhance
emergency healthcare, Journal of Biomedical Discovery andCollaborations 4 (2009) 4.
[11] M.N. Kamel Boulos, D. Burden, Web GIS in practice V: 3-Dinteractive and real-time mapping in Second Life,International Journal of Health Geographics 6 (51) (2007),Available at: http://www.ij-healthgeographics.com/content/pdf/1476-072X-6-51.pdf(last accessed 2.12.09).
[12] M.N. Kamel Boulos, Novel emergency/public health situationrooms and more using 4-D GIS, in: Presented at: ISPRS WGIV/4 International Workshop on Virtual Changing Globe forVisualisation & Analysis (VCGVA2009), Wuhan University,Wuhan, Hubei, China, 2728 October, 2009 (Published inISPRS Archives, vol. XXXVIII ISSN No: 1682-1777 PART4/W10). Available at: http://www.isprs.org/proceedings/XXXVIII/4-W10/papers/VCGVA2009 03608 Boulos.pdf(last
accessed 10.02.10).[13] A. Sanfilippo, A.J. Cowell, L. Malone, R. Riensche, J. Thomas,
S. Unwin, P. Whitney, P.C. Wong, Technosocial predictiveanalytics in support of naturalistic decision making, in:Proceedings of NDM9, the Ninth International Conferenceon Naturalistic Decision Making, London, UK, June, 2009.
[14] NIH RFA-HD-08-023, Innovative Computational andStatistical Methodologies for the Design and Analysis of
Multilevel Studies on Childhood Obesity (R01). Request forApplications (RFA) Number: RFA-HD-08-023, Department ofHealth and Human Services, National Institutes of Health,2008.
[15] NIDA, Drug, Brains and Behavior: The Science of Addiction.National Institutes of Health Publication No. 07-5605,National Institute on Drug Abuse, 2007.
[16] Q. Wiktorowicz, Islamic Activism: A Social Movement
Theory Approach, Indiana University Press, Bloomington,IN, 2004.
[17] A. Gore, An Inconvenient Truth: The Planetary Emergency ofGlobal Warming and What We Can Do About It, RodaleBooks, New York, 2006.
[18] S. Mullainathan, R.H. Thaler, Behavioral economics,International Encyclopedia of the Social & BehavioralSciences (2001) 10941100.
[19] J. Hawkins, S. Blakeslee, On Intelligence, Owl Books, NewYork, NY, 2004.
[20] G. Gigerenzer, Gut Feelings: The Intelligence of theUnconscious, Penguin Books, New York, NY, 2007.
[21] M. Gladwell, Blink: The Power of Thinking WithoutThinking, Little, Brown and Company, Boston, MA, 2005.
[22] G. Klein, Sources of Power: How People Make Decisions, MIT
Press, Cambridge, MA, 1998.[23] I. Janis, Victims of Groupthink: A Psychological Study of
Foreign-policy Decisions and Fiascoes, Houghton, Mifflin,Boston, 1972.
[24] J. Surowiecki, The wisdom of crowds: why the many aresmarter than the few and how collective wisdom shapesbusiness, in: Economies, Societies and Nations, Little,Brown, London, 2004.
[25] G. Miller, The magical number seven, plus or minus two:some limits on our capacity for processing information, ThePsychological Review 63 (1956) 8189.
[26] R.J. Heuer Jr., Psychology of Intelligence Analysis, Center forthe Study of Intelligence, Central Intelligence Agency,Washington, DC, 1999.
[27] A. Tversky, D. Kahneman, The framing of decisions and the
psychology of choice, Science 211 (4481) (1981) 453458.[28] D. Kahneman, A. Tversky, On the psychology of prediction,
Psychological Review 80 (1973) 237251.[29] P. Polgreen, Y. Chen, D. Pennock, F. Nelson, Using internet
searches for influenza surveillance, Clinical InfectiousDisease 47 (November (11)) (2008) 14431448.
http://dx.doi.org/10.1016/j.cmpb.2010.02.007http://doi.acm.org/10.1145/1242572.1242611http://www.ij-healthgeographics.com/content/pdf/1476-072X-6-51.pdfhttp://www.ij-healthgeographics.com/content/pdf/1476-072X-6-51.pdfhttp://www.isprs.org/proceedings/XXXVIII/4-W10/papers/VCGVA2009_03608_Boulos.pdfhttp://www.isprs.org/proceedings/XXXVIII/4-W10/papers/VCGVA2009_03608_Boulos.pdfhttp://www.isprs.org/proceedings/XXXVIII/4-W10/papers/VCGVA2009_03608_Boulos.pdfhttp://www.isprs.org/proceedings/XXXVIII/4-W10/papers/VCGVA2009_03608_Boulos.pdfhttp://www.isprs.org/proceedings/XXXVIII/4-W10/papers/VCGVA2009_03608_Boulos.pdfhttp://www.ij-healthgeographics.com/content/pdf/1476-072X-6-51.pdfhttp://www.ij-healthgeographics.com/content/pdf/1476-072X-6-51.pdfhttp://doi.acm.org/10.1145/1242572.1242611http://dx.doi.org/10.1016/j.cmpb.2010.02.007