Epidemic Alert System: A Web-based Grassroots Model

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 8, No. 5, October 2018, pp. 3809~3828
ISSN: 2088-8708, DOI: 10.11591/ijece.v8i5.pp3809-3828  3809
Journal homepage: http://iaescore.com/journals/index.php/IJECE
Epidemic Alert System: A Web-based Grassroots Model
Etinosa Noma Osaghae1
, Kennedy Okokpujie2
, Charles Ndujiuba3
, Olatunji Okesola4
,
Imhade P. Okokpujie5
1,2,3
Department of Electrical and Information Engineering, Covenant University, Nigeria
4
Department of Computer and Information Science, Covenant University, Nigeria
5
Department of Mechanical Engineering, Covenant University, Nigeria
Article Info ABSTRACT
Article history:
Received May 3, 2018
Revised Aug 20, 2018
Accepted Aug 27, 2018
Most web-based disease surveillance systems that give epidemic alerts are
based on very large and unstructured data from various news sources, social
media and online queries that are parsed by complex algorithms. This has the
tendency to generate results that are so diverse and non-specific. When
considered along with the fact that there are no existing standards for mining
and analyzing data from the internet, the results or decisions reached based
on internet sources have been classified as low-quality. This paper proposes a
web-based grassroots epidemic alert system that is based on data collected
specifically from primary health centers, hospitals and registered
laboratories. It takes a more traditional approach to indicator-based disease
surveillance as a step towards standardizing web-based disease surveillance.
It makes use of a threshold value that is based on the third quartile
(75th
percentile) to determine the need to trigger the alarm for the onset of an
epidemic. It also includes, for deeper analysis, demographic information.
Keyword:
Alert
Disease
Epidemic
Health-care
Infectious
Model
Surveillance
Traditional
Web-based
Copyright © 2018 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
Kennedy Okokpujie,
Departement of Electrical and Information Engineering,
Covenant University,
Ota, Ogun State, Nigeria.
Email: kennedy.okokpujie@covenantuniversity.edu.ng
1. INTRODUCTION
In the last quarter of the year 2017, a rare disease known as Monkey Pox, broke out in Nigeria, a
nation in the western part of Africa. The information about the disease outbreak got to the Nigeria Centre for
Disease Control (NCDC) through the Niger Delta University Teaching Hospital (NDUTH), Okolobiri,
Bayelsa State. By the end of the year, a total of forty three (43) cases spread across eight states in Nigeria
have been confirmed. New cases of the Monkey Pox disease stopped being reported in Nigeria by the start of
the year 2018. This was quite laudable as Nigeria uses traditional surveillance methods to watch out for
disease outbreaks [1]. This case study reinforces the need to detect the outbreak of infectious diseases at the
earliest stage, especially at this time when the world has become a “global village”. Thus, nations invest
heavily on disease surveillance systems. The reason is not far-fetched, the outbreak of an infectious disease,
if not contained at its earliest stage, could lead to catastrophic local, national and world-wide consequences.
Economies can be brought to their knees by epidemics that could not be contained because it was not
detected early.
The internet has become a powerful tool for detecting epidemics at its earliest stage, it has made it
possible to collate and deliver information on the progression of disease outbreaks, epidemics and in some
cases, pandemics within days or even hours. The power of the internet is now being explored on a worldwide
scale for disease surveillance. A lot of attention has been shifting lately to the possibilities embedded in the
internet for web-based disease surveillance [2]. Today, a lot of web-based systems serve the world in various

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 8, No. 5, October 2018 : 3809 – 3828
3810
languages and utilize data from news sources and social media to detect epidemics at their earliest stage.
These web-based disease surveillance systems are sometimes restricted, semi-restricted or freely accessible
to the public.
Event-based surveillance systems usually utilize data from online sources. The data acquired could
be moderated or aggregated automatically. Syndromic-based surveillance and indicator-based surveillance
system make use of health data from healthcare providers, diagnostic laboratories and surveillance specialists
in governmental organizations. There also exists a whole arsenal of web-based disease surveillance systems
that give early alerts about the outbreak of diseases based on queries made by internet users. The main goal
of all these web-based systems remains the early detection of an epidemic outbreak.
The United States can boast of Program for Monitoring Emerging Diseases (ProMED-mail) and
Epi-SPIDER for web-based disease surveillance and bio-security intelligence. Other well-known disease
surveillance systems like Influenzanet (Europe), The Global, Public Health Intelligence Network - GPHIN
(Canada), The Global Outbreak Alert and Response Network (GOARN) and Google Flu Trends are web-
based. These web-based epidemic alert systems all make use of data mined from the internet and complex
algorithms to analyze data for useful information on disease outbreaks. But most of these systems extract
information from a large pool of data on a very large number of infections and thus have difficulty presenting
critical information concisely and without ambiguities. The reason for this is simply the fact that till date, no
sound methodology has been developed for measuring the relationship between data mined on health related
issues from the internet and actual public health related issues like epidemics and pandemics.
In this paper, a shift to the use of a more traditional approach that is web-based is proposed. It would
make use of clinical data from primary health centers, diagnostic laboratories and hospitals. The data
collected would be classified using the syndromic codes contained in the tenth revision of The International
Classification of Diseases (ICD-10). A weekly percentile check forms the basis for determining if a disease
has reached epidemic levels. Algorithms working at the background analyses information got from a network
of primary health centers, hospitals and laboratories to generate graphically illustrated results about fifty (50)
diseases including those on the World Health Organization (WHO) watch list. A threshold based on the third
quartile (75th Percentile) for each week was to determine when to trigger the epidemic alert. Diseases on the
watch list, when detected, trigger a special alarm. Information like demographics and location were also
included for more detailed analysis.
2. LITERATURE REVIEW
2.1. The big argument
The big argument about the true difference between the words, Epidemic and Pandemic was
addressed by David M. Morens et al [3]. In a bid to describe pandemics, the researchers elucidated on eight
key factors that tend to characterize a widespread infection as a pandemic. In the expository, the researchers
surmised that a disease can be termed pandemic if:
a. It covers a very large geographic area. Trans regional (greater than or equal to two adjacent regions of the
world); inter regional (greater than or equal to two non-adjacent regions) and global.
b. It can be traced from place to place as it progresses.
c. It spreads explosively and has a high attack rate.
d. The population has a minimal immunity to it.
e. If it is novel (like HIV/AIDS), through a new strain of the same pathogen.
f. If it is very contagious without specific regards to the means by which it is spread.
g. If it is very severe.
In all, the authors concluded that having the term Pandemic, defined as a large epidemic makes
sense and avoids the pitfalls of inconsistency. In addition, the researchers suggested that the term pandemic
be used for only infectious diseases.
2.2. Web-based disease surveillance and epidemic alert system: state-of-the-art
Erini reviewed some of the latest technologies and tools used to carry out regional and global
infectious disease surveillance [4]. A review on epidemic modeling was also done. The need to be able to
quickly and efficiently classify previously unknown strains of pathogens that are responsible for emerging
infectious diseases was stressed. Disease surveillance with the latest and most effective tools was also
encouraged to ensure that novel and re-emerging infectious diseases do not attain epidemic/pandemic levels.
The author emphasized the need for large-scale infectious disease surveillance networks, especially in the
world of today that is fast becoming smaller due to “never-seen-before” bridges in communication and
transportation gaps. The author highlighted event-based surveillance, web-based real-time surveillance, early

Int J Elec & Comp Eng ISSN: 2088-8708 
Epidemic Alert System: A Web-based Grassroots Model (Etinosa Noma Osaghae)
3811
warning and alert response networks, infectious diseases modeling, social media and new technologies in
pathogen discovery as the key drivers in the new age of disease surveillance.
Collier, of the National Institute of Informatics, Tokyo, Japan did a survey on the importance of
Epidemic Intelligence (EI). The focus of the researcher‟s work was a survey on the use of artificial
intelligence, social network, data mining tools and natural language processors to monitor the progress of
disease outbreaks from predominantly unstructured data. He surmised that at the core of Epidemic
Intelligence using unstructured data was the technology called Text Mining [5].
In their work, Jennifer L. Gardy et al shed light on the need to develop a disease surveillance system
that is global in scope, works on the go and is genomics-informed especially iafter the Ebola and Zika
epidemics. The researchers proposed a One Health System that is based on genomics diagnostics and
epidemiology integration into existing disease surveillance systems [6]. The researchers envisioned a system
that integrates human, animal and environmental health to proffer disease surveillance solutions to regions of
the world with inadequate to non-existent laboratory facilities [7]. The researchers paid particular attention to
how several epidemics like the Ebola and Zika progressed uninhibited for months and unnoticed by even the
most advanced disease surveillance systems until they were discovered much later when they have grown to
epic proportions. Novel pathogen identification and the detection of certain old pathogens in new regions
were identified as the major gaps in all existing disease surveillance systems. Major impediments to the One
Health Scheme proposed by the researchers were identified as government policies, conflicts between
medical practitioners and researchers from lower income, middle income and better resourced nations. The
future of disease surveillance as highlighted by the researchers is the genomics-informed [8] one where all
the factors affecting health are considered on a global scale with complete and uninhibited transparency.
S.J. Yan et al explored the accuracy and timeliness of data mined from unstructured sources on the
internet for Epidemic Intelligence (EI). The researchers surveyed a lot of publications on the subject of Text
Mining for Epidemic Intelligence and came up with the conclusion that serious attention should be paid to
the timeliness and accuracy of information about disease outbreaks got from mining ubiquitous, free and
unstructured data from the internet [9].
Simon Pollet et al focused on the use of “Big Data” to get Epidemic Intelligence (EI) on vector-
borne diseases (VBD) in middle and low income countries [10]. The research was a review on the
performance of various internet-based tools and techniques that have been employed to mine data on vector-
borne disease. The researchers carried out an in-depth survey and came to the conclusion that more reviews
need to be done to ascertain the true impact of using “Digital Epidemiology” in tandem with more
conventional or traditional means of disease surveillance. The research also called for more surveys on the
reaction of end users to the metrics used to classify or gauge the outbreak of a disease. The researchers also
emphasized that “Digital Epidemiology” was not made to replace but rather complement traditional methods
of vector-borne disease surveillance.
Eun Kyong Shin et al looked closely at the progress of online clinical trials in the United States of
America. The research focused on the popularity and impact of online clinical trials and health studies from
the first time it appeared online and its perceived future. The research work also exhaustively detailed the
potential and obvious use of the internet for health studies.
Natalie S. and Collins A. in the article, Web-based Surveillance of Illness in Childcare Centers,
made a proposal for active bio-surveillance in childcare centers [11]. The authors were of the opinion that
monitoring childcare centers for disease outbreak was more effective than the traditional method of
monitoring schools only. The article pointed out summer breaks as one of the main or core reasons why
monitoring schools only for disease outbreaks was not effective enough. The authors also proposed that the
bio-surveillance of childcare centers be web-based and should submit reports on a weekly basis to the central
public health department. Some of the key metrics or data that was monitored by the proposed child care bio-
surveillance program were children categories (toddlers, infants and pre-schoolers) and the illness reported.
The statistical report was expressed in terms of percentages and actual whole numbers for each major
category. The authors claimed that the system, implemented in a Michigan County (United States of
America), was able to detect the outbreak of Gastroenteritis and Hand-Foot-Mouth disease when the more
conventional school based disease monitoring system was not available, especially during winter and summer
seasons.
The Global Public Health Intelligence Network (GPHIN) was credited with sending the first alert on
the Acute Respiratory Illness Outbreak, code-named, MERS-CoV (Middle East Respiratory Syndrome
Coronavirus) [12]. It is a web-based program that uses specialized algorithms to harness the power of Big
Data to mine for clues that signal the onset of an epidemic. The web-based program, in conjunction with a
multilingual and multidisciplinary team, culled and analyzed information from over thirty thousand sources
in nine languages for potential clues to the onset of an epidemic anywhere in the world. The authors made it
known that the system is being adopted by many nations for national disease surveillance. The authors

 ISSN: 2088-8708
3812
emphasized how future GPHIN projects plan to utilize more of the power of Big Data especially from social
media outlets using sophisticated algorithms to mine for epidemic clues.
The use of drivers of emerging infectious diseases [13], [14] was suggested by Sarah H. Olson et al
to develop the framework for digital detection of Infectious Disease (ID) events. The researchers were of the
opinion that close monitoring of infectious disease drivers could provide a viable means for the early
detection of potential infectious disease epidemics especially in the case of emerging infectious diseases. The
researchers identified some of these drivers as climate and meteorological data. The researchers also
presented a sample framework for the use of Infectious Disease (ID) drivers in digital disease surveillance
programs. An extensive review of previous infectious driver models was also done and the gaps were
identified.
Nsoesie et al carried out an extensive review of the most recent digital technologies that have been
employed for infectious disease surveillance at mass gathering events. Interestingly, among the digital
technologies was the internet or web-based approach to disease surveillance at mass gathering event [15].
Notable among the web-based digital disease surveillance for mass gathering events was the Healthcare
Electronic Surveillance Network (HESN) implemented by the Kingdom of Saudi Arabia to closely monitor
respiratory, gastrointestinal, cardiovascular, skin and ear/nose diagnoses. The system was said to have been
very effective during the 2013 Hajj season. The HESN captures diagnostic data from healthcare practitioners,
clinic and hospital staff, paramedics and other health related outfits for semi-automatic analysis and prompt
decision taking. In the 2002 Salt Lake City winter games, a web-based and fully automated infectious disease
surveillance system was used to analyze health data from several sources and most prominent was the
triggering of an alert if any disease outbreak is suspected. The system actually gave two alerts for respiratory
infections that were promptly put addressed by health officials. The impact of web-based epidemic alert
systems during many other mass gathering events like the world cup and religious gatherings were
elucidated. Small gatherings were not left out. A good example given was the use of a combo of the Global
Public Health Intelligence Network (GPHIN) and Medical System (Medisys), all web-based disease
surveillance systems, for the infectious disease surveillance during the 2012 European football championship.
Some researchers [16] carried out a systematic review to determine the extent and depth to which
Online Social Networks (OSNs) have been used for disease surveillance. The study submitted that lots of
models, framework and systems for disease surveillance using online social networks have been developed
and in many cases, implemented. The researchers acknowledged the fact that online social media provides a
viable means of tracking pandemics because of its vast and varied, though unstructured nature. The large
population of people from various places all-over the world and the exchange of information that goes on
unabated via social media platforms have been mined and analyzed by an array of complex algorithms and
computational linguistics to track pandemics. The criteria used for each OSN pandemic tracking system were
numerous and made it clear that the use of online social networks to track the onset or progress of a pandemic
may never replace traditional and more conventional methods of disease surveillance.
The need for web-based disease surveillance is also being explored and implemented by the armed
forces as seen in the joint bio-surveillance portal championed by the Republic of Korea and the United States
of America [17].
The use of purely traditional means of disease surveillance and the emergence of the WHO as the
international instrument for the expediting of intervention programs in the event of unusual and especially
tough epidemics and pandemics has led to the disappearance of a number of infectious diseases around the
world [18]. But emphasis today keeps shifting towards real-time disease surveillance [19]. The report
submitted that the state-of-the-art for real-time disease surveillance depends heavily on social media.
Jihye choi et al also reviewed various web-based infectious disease surveillance systems that have
been used. The study focused on the current state-of-the –art, benefits and challenges associated with web-
based disease surveillance systems that have been implemented in various ways to support the more
conventional or traditional surveillance methods [20]. The authors looked closely at the strengths and
weaknesses of eleven web-based surveillance systems and gave comments on how the weakness of some of
the web-based surveillance methods already in use, can be improved. The researchers did submit that web-
based disease surveillance methods were adaptable, low-cost, and intuitive and can be operated in real-time.
The researchers identified privacy issues, prediction and an interpretation inaccuracy as some of the potential
challenges of internet-based epidemic alert systems. The authors also noted the absence of a functional web-
based epidemic monitoring systems in some nations with advanced information and communication
technology presence. The authors also classified standard disease surveillance systems and a lot of them were
web-based. This is shown in Figure 1.

3813
Figure 1. Classification of standard disease surveillance systems [20]
A research work carried out during a project backed by the European Commission, acknowledged
the high outlay of capital on electronic disease surveillance systems to help in detecting the outbreak of
emerging and re-emerging infections on time. The report submitted that it remains unclear if existing and
sophisticated real-time electronic surveillance systems can effectively detect the outbreak of an epidemic
early.
The Sustainable Surveillance Workgroup made some suggestions on how to build a sustainable
disease surveillance system [22] that is equipped to provide information about infection outbreak
continuously. The report supported the need to know that continuous and unabated disease surveillance is a
must for the benefit of public health. The report also stressed the need to improve on the monetary allotment
for surveillance purpose, have an active surveillance workforce and delve into deep rigorous disease
surveillance research that would lead to a better understanding of public health and help with the creation of
policies and decision making.
The lessons learned from the various implementations of web-based disease surveillance systems
can be seen in the article written by MO Lwin et al of how Mo-Buzz [23], a mobile pandemic surveillance
system for Dengue was implemented in Colombo, Sri Lanka. The mobile application was developed to take
advantage of Sri Lanka‟s large mobile device using population. The study submitted that the traditional
Dengue reporting structure in Sri Lanka was excruciatingly slow because it was still paper based. The
introduction of Mo-Buzz in two phases, one for the general public and another for Sri Lanka‟s health
institutions, led to a boost in the country‟s ability to detect, keep track and inform the public about Dengue
disease outbreaks. The researchers noted that though Mo-Buzz‟s initial uptake was quite low, it picked up
and went as high as 76%. This study confirms the fact that mobile and social media outlets which are all
web-based are the future for global disease surveillance [24].
2.3. Traditional and syndromic surveillance of infectious diseases and pathogens
Cedric Abat et al noted that many disease surveillance systems are in use all around the world [25].
The reasearchers made a summary of some disease surveillance methods in use all around the world. The
researchers looked at syndromic surveillance from the microbiology perspective. The researchers submitted
that disease surveillance data can be gathered from the Human Environment with focus on Environmental
data (water pollution, weather, and air pollution) and Animal Health data (information about the health of
domestic and wild animals). Surveillance data can also be got from human behaviour, which consists of
Internet use (web queries, press dispatches, social media, press articles), Telephone (hotlines), Drug sales and
Absenteeism. Health Care also provide viable disease surveillance data via Sentinel surveillance (sentinel
physicians who agree to notify the public health authorities at regular intervals of patients presenting certain

 ISSN: 2088-8708
3814
specific symptoms of infectious diseases on the watch list), Chief complaints, Medical records, Hospital
discharge data, Microbiology orders, Disease reports and Demographics.
The researchers listed some surveillance strategies such as Disease-specific surveillance, Event-
based surveillance and Syndromic surveillance [26].
2.4. Infodemiology metrics
Infodemiology helps deternine the best way to tackle the issue of public health. It is possible to
collect data for this information in real-time. Internet queries have been used to predict the outbreak of an
epidemic. Twitter microblogs, the news and the way people use the internet for health services have been
monitored. The information gleaned from all these numerous sources are analysed and useful information
that can be used to inform about health policies are inferred [27]. A lot of metrics have been used to gauge
the impact of information got from the internet.
It has been said that there is a need to standardize infodemiology and inforsurveillance metrics.
Infodemiology is primarily and electronic (got using some kind of algorithm). According to the authors,
infodemiology‟s most basic metrics could be supply related (internet users postings) or demand related
(internet users buying habits) [28]. On the supply side, the most basic metrics were information prevalence
and information occurrence ratios. On the demand side, the most basic metrics were the number of searches
of a specific topic and number of clicks on a website about a specific topic. An active method involving
online surveys of consumers of health products also provides for a good infodemiology metric.
2.5. Models for web-based surveillance and epidemic alert systems
The ubiquitous and pervasive powers of the internet and social media have forced many states to
revise their disease surveillance policies. It is almost a must for states to carry out active disease surveillance
and inform the World Health Organization (WHO) of any epidemic outbreak. Individuals and non-
governmental organizations have continued unabatedly to use the power of social media, the internet and
complex algorithms to report cases of epidemic outbreaks to WHO and in many instances, before it is known
and accepted officially [29]. The model is shown in Figure 2.
Figure 2. Decision-making instrument for International Health Regulation - IHR (2005), adapted from
Annex 2 [29]
WHO has an integrated global alert and response system for epidemics and pandemics. The system
is based on already existing and effective national health systems and an international coordinated response
system. Presently, WHO get alerts about epidemics and pandemics through the health care system of its
member states? The model uses phases of this nature for zoonotic diseases;

3815
Phase 1-3: The phase for preparing, gathering up-to-date information and planning for emergency
response.
Phase 4-6: Actual emergency response and mitigating efforts.
In a more detailed format, the various phases have these components:
Phase 1: No human infections. Only animals are affected.
Phase 2: Humans have been infected but at very low to non-existent levels.
Phase 3: A tangible number of humans infections have been recorded but no human-human
transmissions yet.
Phase 4: Verified and recorded human-human infection with attendant “community-level outbreaks”
or full epidemic.
Phase 5: Verified and recorded human-human infection with attendant “community-level outbreaks”
in at least two countries within a defined WHO region. Pandemic is very imminent.
Phase 6: Verified and recorded human-human infection with attendant “community level outbreaks”
in at least two countries within a defined WHO region and at least one country in a different WHO region. A
pandemic is underway.
Presently WHO has made it necessary for all national governments to report cases of any disease on
a watch list that keeps growing with new additions.
A Bayesian Hierarchical Poisson Model with a hidden Markov model was proposed by
D. Conesa et al for the early detection of influenza epidemic outbreak [30]. The model relied greatly on an
intensity parameter that was set by the incidence frequency. The incidence rate was considered as a normal
distribution in which its parameters, mean and variance, were modeled to reflect the phase of the system, be
it epidemic or non-epidemic. The transition took into cognizance previous weekly epidemic states. The
authors gave samples of how to implement the statistical model and used Bayesian Inference to define the
state of an influenza epidemic at any moment. The researchers gave the transition probabilities as:
( ) * +
Where;
An observed random variable that indicates the phase of the modeled system as either
epidemic ( ) or non-epidemic ( )
Suitable probabilities
Day (during the week)
Season
The Moving Epidemic Method (MEM) was used to model the incidence rate of influenza-like
illness (ILI) and Acute Respiratory Illness (ARI) for some European countries by [31]. The values obtained
were used to compute the various intensity levels adopted by the research, namely, Baseline, Low, Medium,
High and very High. The researchers used these benchmarks to compare the epidemic level of ILI in various
European nations for different time periods from 1996/1997 – 2013/2014 seasons. The authors arrived at the
conclusion that these comparisons are important for firm understanding of seasonal epidemic patterns and
thus, should be incorporated into automated disease surveillance systems at national and international levels.
In the article, Zika Virus: A New Pandemic Threat [32], allusion was made to a special software
application, Zika Tracker, that was used to aid voluntary reporting of confirmed Zika virus infection cases to
help the Americas contain the spread of the virus which was suspected to be the main cause of an alarming
rise in Microcephaly cases in the Americas. This is one of the very basic examples of a disease alert and
monitoring system that exploits disease surveillance at the grassroots.
Ruth A. Ashton et al, gave an insightful expository into the usefulness of school-based disease
surveillance with malaria as a case study. [33]. The research focused on a pilot programme that was carried
out in Ethiopia to monitor malaria epidemics and focused particularly on school absenteeism and febrile
illness. The researchers submitted that a lot of challenges hampered the study. The focus on a school-based
system brought a serious challenge of population representation as almost 46% of Ethiopian school aged
children are not enrolled in school. The researchers suggested that another pilot project be carried out again
when there is a substantial increase in reported cases of malaria than what is conventionally known. In all, the
researchers noted that the sensitivity of the school-based syndromic surveillance to detect epidemics could
not be fully ascertained. The Model used is shown in Figure 3.
Four critical areas that pose serious problems to disease surveillance on a global scale were outlined
by [34] as scientific methods, international policies, technical resources, financial resources and human
resources. The researchers gave a notional scheme for a global disease surveillance and response process.

 ISSN: 2088-8708
3816
Some researchers [35] wrote on how epiDMS: Data Management and Analytics for Decision
Making from Epidemic Spread Simulation Ensembles, have helped to plug some critical holes that has to do
with scalability, multiple interdependency parameters and complex dynamic processes during an ongoing
epidemic. The researchers claimed that the data management and analytics tools offered by epiDMS help
with the decision-making process in the event of an epidemic with significant health and economic benefits.
Figure 4 depicts the epiDMS model.
Figure 3. Model for school-based malaria epidemic surveillance system [33]
Figure 4. Overview of the epiDMS system [35]
A group of researchers developed an internet-based epidemic alert system for periodontal disease in
Nigeria. The web-based model proposed by the researchers was based on real-time statistical data for
periodontal disease diagnosis across Nigeria. HTML, PHP and CSS were used to develop the user-friendly
interface of the system and MySQL [36] was used to create the database of the system. The researchers
claimed that the proposed system will help with the surveying and tracking of periodontal disease in
Nigeria [37].
Some researchers[38] used colour code based on the alert phases already defined by WHO to
determine and raise an alarm for the outbreak of the AH1N1 Influenza in America. The authors used the
Basic Reproduction number to know when there is a need to trigger an epidemic alert. The authors
utilized accumulated data from sixteen (16) out of the thirty five (35) member states of the Americas to get
the Basic Representation number. A Basic Representation number greater than one (>1) was the trigger for
the outbreak of an epidemic.

3817
3. EPIDEMIC ALERT THRESHOLD ALGORITHMS
3.1. The use of change point analysis
A change point is the point where a structure change occurs in the collected data [39]-[ 40]. The
series can be represented as:
{ }
And the index of time, represented as:
* +
The epidemic or endemic component of the process is a piecewise constant. The pre-epidemic
period (endemic state), epidemic period (epidemic state) and the post-epidemic (endemic state) would be
determined. A changes are first detected, counted and estimate. If * + is the time series of
independent variables and , where represent the corresponding structure parameters then, a
decision has to be made between:
No change point
And
Change points
Note:
a.
b. represents the start and end dates of the outbreak respectively
c. The rejection of confirms a change point.
If is rejected, the number of changes in state and their actual position has to be estimated. Thus:
If is true, what is and from the sample, { }. This change point problem was
solved using the non-parametric kernel model. Based on simulated data, the non-parametric kernel model
was used to detect the start of an outbreak and the end of an outbreak.
3.2. The kernel model
If { } is a true series of independent random variables, the Kernel function is defined as:
( ) * +
A Kernel Fisher discriminant ratio (KFDR) is used to measure the heterogeneity between successive
segments, .
{ } with observations, pre-epidemic.
{ } with ( ) observations; epidemic.
* + with ( ) observations; post-epidemic.
A simple linear kernel function:
( ) is used to determine the value of
To find the KFDR between and :
( )
( ) ( )
( )
( ) ( )
are chosen to maximize the heterogeneity between the three segments by calculating:
( )
( )
( )
( )( )
( )
( ̂ ) ( ) * + * + ( ( ))

 ISSN: 2088-8708
3818
4. RESEARCH METHOD
The web-based grassroots model assumes a medical history and uses the third quartile value for
each disease under consideration to set the threshold that would determine when a disease has reached
epidemic levels. The proposed system uses the weekly percentile.
The World Health Organization suggested the use of the third quartile (75th
Percentile) as a
threshold for triggering the onset of an epidemic [41].
[ ]
The percentile
Limit below the the desired percentile point‟s interval
Total available scores
The score point in terms of desired percentile
Summation of frequency scores below the percentile point‟s interval
The percentile poin‟s frequency scores
Class interval width
A baseline or threshold is set, beyond which a disease has reached epidemic levels. The „Low”
alarm is triggered when the seventy fifth percentile value (third quartile) for a given disease is higher than the
baseline at any given week. The “Moderate” alarm is triggered when the third quartile value of a given
disease is higher than the baseline for two consecutive weeks. The “High” alarm is triggered when the third
quartile value of a given disease is higher than the baseline for three consecutive weeks. The “Severe” alarm
is triggered when the third quartile value of a given disease is higher than the threshold for four consecutive
weeks.
On the user side, data collected using the proposed web-based grassroots model includes:
1. Medical: Centre Code, Age, Sex, Symptoms, Diagnosis (50 ICD codes were used in this demonstration).
2. Microbiology Orders: Laboratory Centre code and Pathogen.
3. Notifiable Disease Report: Express Notification by Sentinel Surveillance.
On the administrator side, data monitored include:
1. Dashboard: Special Alerts and Express Notification by Sentinel Surveillance.
2. Percentile graphs: Daily and Weekly Percentile Graphs.
The International Classification of Diseases (ICD-10) codes and their respective threshold based on
previously known (assumed in this case) weekly third quartile values for the sample diseases monitored by
the proposed web-based grassroots model are shown in Table 1.
Table 1. ICD Codes and their Third Quartile Baseline
S/N Disease ICD Code Third Quartile
Baseline
1 Cholera* A00 5
2 Plague* A20 0
3 Yellow Fever* A95 0
4 Small Pox* B03 0
5 Relapsing Fever* A68 0
6 Typhus* A75 15
7 Polio* A80 0
8
Severe Acute
Respiratory
Syndrome
(SARS)*
J60 30
9
Ebola virus
disease*
A98-4 0
10 Influenza* J10 0
11 Lassa Fever* A96-2 0
12
Marburg
Hemorrhage
Fever*
A98-3 0
13 Rift Valley Fever* A92-4 0
14 Tularemia* A21 0
15
Dengue
Hemorrhagic
Fever*
A91 0

3819
S/N Disease ICD Code Third Quartile
Baseline
16
Crimean-Congo
Hemorrhagic
Fever*
A98-0 0
17 Anthrax* A22 0
18 Monkeypox B04 12
19 Candidiasis B37 10
20 HIV/AIDS B20 20
21 Diarrhea A09 30
22 Tuberculosis A16 50
23 Rabies* A82 0
24 Botulism A05-1 23
25 Campylobacteriosis A04-5 15
26 Chickenpox B01 35
27
Creutzfeldt-Jakob
Disease
A81-0 20
28 Dysentery A06-0 12
29
Hantavirus
Pulmonary
Syndrome
I26 11
30 Helicobacter Pylori K31-2 5
31 Hepatitis B B16 12
32 Hepatitis C B17-1 12
33 Histoplamosis B39 2
34 Leptospirosis A27 35
35 Lyme Disease A69-2 13
36 Measles B05 15
37 Mumps B26 15
38
Typhoid and
Paratyphoid Fevers
A01 70
39 Diptheria A36 48
40 Schistosomiasis B65 50
41 Tetanus A33 10
42 Taxoplasmosis B58 25
43 Leprosy A30 10
44 Viral meningitis A87 10
45 West Nile Virus A92-3 10
46 Dyspepsia K30 25
47 Hepatitis A B15 60
48 Whooping Cough A37 5
49 Malaria B50 100
50 Scabies B86 35
*The asterisked diseases on a special watch list
NOTE: The International Statistical Classification of Diseases (ICD) – 10 codes used herein does not take into
cognizance subsets of the code for disease variations and causative organisms.
For demographic analysis, codes were assigned to hospitals, primary health centres and laboratories that
provide inputs to the epidemic alert system. Table 2 shows some sample centre code for hospitals,
laboratories and primary health centres that were used to demonstrate how the proposed epidemic alert model
works. Table 3 shows a sample of the inputs got from Laboratories. Table 4 shows a sample of the inputs got
from medical records in hospitals and primary health centres. Figure 7 shows the algorithm for the proposed
web-based grassroots epidemic alert system.
Table 2. Centre Codes
S/N Centre Code Type Location
1 H01865360875 Hospital Marque, Kingston
2 P12765098656 Primary Health Centre Dale, Lofty Heights
3 P37659339059 Primary Health Centre Tomahawk, Prowess
4 L78599539584 Laboratory Balinese, Catwalk
5 H24698736483 Hospital Cross, Time Hills
6 H54786783995 Hospital Action Yard, Trent
7 P84898479948 Primary Health Centre Hague, Bella vane
8 H67494672997 Hospital
Seminary Zone,
Primer
9 L75643782674 Laboratory Hebron, Simile
10 P98573652641 Primary Health Centre Bayville, Manama

 ISSN: 2088-8708
3820
Table 3. Microbiology Orders
Centre Code Pathogen
Test
counts
H01865360875
Vibro Cholerae Plasmodium
Spp
Salmonella Spp
5
13
6
P12765098656
Spp
Salmonella Spp
3
15
18
P37659339059
Spp
Salmonella Spp
6
45
12
L78599539584
Vibro Cholerae
Plasmodium Spp
Salmonella Spp
4
20
7
H24698736483
Vibro Cholerae
Plasmodium Spp
Salmonella Spp
8
34
9
H54786783995
Vibro Cholerae
Plasmodium Spp
Salmonella Spp
2
12
5
P84898479948
Spp
Salmonella Spp
7
60
1
H67494672997
Spp
Salmonella Spp
2
31
4
L75643782674
Vibro Cholerae
Plasmodium Spp
Salmonella Spp
8
15
12
P98573652641
Vibro Cholerae
Plasmodium Spp
Salmonella Spp
6
9
4
Table 4. Medical Records
Centre Code Age Sex
Diagnosis (ICD-Code)
- Counts
H01865360875 0 – 1 (Infant)
1 – 4 (toddler)
5 – 12 (child)
Male
Female
Male
Female
Male
Female
B50 – 23, A01-4, A00
– 21, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2,
B50 – 5, A00 – 0, B16
– 7, B37 – 1.B50 – 23,
A01-3, A00 – 12, B05
– 0, B01-2, A09 -0,
A80 – 3, B86 – 2, B50
– 5, A00 – 0, B16 – 1,
B37 – 0.
B50 – 23, A01-7, A00
– 21, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 0,
B16 – 7, B37 – 0.
B50 – 22, A01-4, A00
– 21, B05 – 1, B01-3,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 0,
B16 – 7, B37 – 0.
B50 – 15, A01-4, A00
– 21, B05 – 1, B01-2,
A09 -1, A80 – 5, B86
– 2, B50 – 5, A00 – 5,
B16 – 7, B37 – 1.
B50 – 20, A01-4, A00

3821
Centre Code Age Sex
Diagnosis (ICD-Code)
- Counts
13 – 17 (Teenager)
18 – 59 (Adult)
60 and above
(elder)
Male
Female
Male
Female
Male
Female
– 21, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 1,
B16 – 7, B37 – 1.
B50 – 22, A01-4, A00
– 21, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 5,
B16 – 7, B37 – 1
B50 – 21, A01-4, A00
– 18, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 3,
B16 – 7, B37 – 0.
B50 – 18, A01-4, A00
– 20, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 4,
B16 – 7, B37 – 0.
B50 – 15, A01-4, A00
– 16, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 2,
B16 – 7, B37 – 1.
B50 – 20, A01-4, A00
– 15, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 2, B50 – 5, A00 – 4,
B16 – 5, B37 – 1.
B50 – 20, A01-4, A00
– 14, B05 – 1, B01-2,
A09 -1, A80 – 3, B86
– 1, B50 – 5, A00 – 2,
B16 – 7, B37 – 0.
4.1. The model for the user interface
Figure 5 shows the model of the user interface for the proposed web-based grassroots epidemic alert
system.
4.2. Model for administrator interface
Figure 6 shows the model of the administration interface for the proposed web-based grassroots
epidemic alert system.

 ISSN: 2088-8708
3822
Graphical User Interface
Medical Records Microbiology Orders
Notifiable Disease
Report
Centre Code Age Sex ICD Code
Centre Code Pathogen
Diseases on
Watch List
Undiagnosed
Diseases
Database
Figure 5. Model of the user interface for the proposed web-based grassroots epidemic alert system
Epidemic Alerts
Special Alerts for
Diseases on Watch List
Alert Level:
Low
Alert Level:
Moderate
Percentile
Graphs
Bar Charts
Demographics
Location Sex
Causative
Organism
Undiagnosed
Diseases
Investigate Further
Alert Level:
High
Alert Level:
Severe
Age
Figure 6. Model of the administration interface for the proposed web-based grassroots epidemic alert
system
5. RESULTS AND ANALYSIS
The baseline values for all diseases being monitored is depicted in Figure 8 and Figure 9. It is based
on the corresponding third quartile value for each ICD code. Some ICD codes like A20, A82, and A80 have
corresponding baseline or threshold values of zero. This is because they are diseases on the special watch list.

3823
It is assumed that any recorded case of the diseases on the watch list should generate swift and appropriate
action to protect public health.
In Figure 10, is shown the minimum, first quartile, median, third quartile and maximum values for
ICD codes, A01, A36, B65, A33 and B58. It can be seen clearly from the plot that none of the diseases under
observation has exceeded their respective threshold values. A close look at the plot reveals that the third
quartile value for the observed diseases can be shown in a tabular format as shown in Table 5.
START
FIRST WEEK:
IS DISEASE CASES HIGHER
THAN THE KNOWN
THRESHOLD
NO
IS DISEASE ON THE
WATCHLIST?
SPECIAL ALERT
NO NON-EPIDEMIC
YES
ALERT: LOW
ALERT: MODERATE
ALERT: HIGH
SECOND WEEK:
THAN THE KNOWN
THRESHOLD
THIRD WEEK:
THAN THE KNOWN
THRESHOLD
FOURTH WEEK:
THAN THE KNOWN
THRESHOLD
ALERT: SEVERE
YES
YES
YES
YES
NO
NO
Figure 7. Algorithm for the proposed web-based grassroots epidemic alert system

 ISSN: 2088-8708
3824
Figure 8. Baseline graph for each ICD code at
the 75th
percentile (third quartile)
Figure 9. Bar representation of weekly
thresholds for all fifty (50) diseases
Figure 10. Weekly baseline graph for selected diseases
The actual third quartile values in Table 5 are all less than the baseline values. Therefore, the
observed diseases are all non-epidemic.
Table 5. Comparison between Actual Third Quartile Values and Baseline Values
ICD
Code
Actual Third
Quartile Value
Baseline Value Epidemic Alerts
A01 60 70 Non-Epidemic
B58 14 25 Non-Epidemic
B58 14 25 Non-Epidemic
In Figure 11 there is an outlier. The outlier corresponds to the ICD code, B50. The outlier has a third
quartile value that is greater than the threshold. But this is not to say that a disease has to be an outlier before
the epidemic alarm is triggered. The only determining factor for triggering the epidemic alert system is the
disease‟s third quartile value at the end of every week. From Table 6 it can be seen that Malaria (B50) has
reached epidemic levels and a “Low Alert” was triggered

3825
Figure 11. Weekly series graph with malaria (B50) at the “low alert” level
Table 6. Malaria (B50) at Epidemic Levels
Disease
Actual Third
Quartile Value
Baseline Value Epidemic Alert
K30 18 25 Non-epidemic
B15 57 60 Non-epidemic
A37 2 5 Non-epidemic
B50 150 100 Low Alert
B86 20 35 Non-epidemic
From Figure 12, it can be observed that the third quartile values of A01 for three consecutive weeks were
above the threshold (baseline) value of seventy (70). This triggered the “High Alert” for Typhoid and
Paratyphoid Fevers. Table 7 shows this result clearly.
Figure 12. Series graph with typhoid (A01) at the “high alert” level
Table 7. Typhoid and Paratyphoid Fevers (A01) at High Alert
Disease Week 1 Week 2 Week 3
A01 80 120 112
In Figure 13, the ICD code A00 has third quartile values that exceed the baseline for four
consecutive weeks. Thus, the “Severe” alarm is triggered for ICD Code, A00. ICD code, A01, though having
third quartile values that are above the threshold for three consecutive weeks, its epidemic status has been
changed to non-epidemic because its third quartile value at the fourth week is less than the threshold value.
This submission is clearly shown in Table 8.
A special exception is made for diseases that are on the watch list. In Figure 14, A80 has a third
quartile value of zero (0) but a maximum value of one (1). Being on the watch list, A80 with a maximum
value of one (1) triggered a “Special” alarm made for diseases on the watch list or diseases that have been
confirmed eradicated officially. Table 9 sheds light on this special case.
A demographic analysis gives the percentage of infants, toddlers, children, teenagers, adults and
aged persons affected by an epidemic outbreak [42]. It also provides information about the location of the
outbreaks. It may also give an idea of the causative organism. A sample of the demographic analysis of
Centre Code, H0186536085 for Cholera outbreak is shown in Figure 15.

 ISSN: 2088-8708
3826
Figure 13. Series graph with cholera (A00) at the “severe alert” level
Table 8. Cholera (A00) and Typhoid Fever (A01) on “Severe” and Non-epidemic Alerts Respectively
Disease Week 1 Week 2 Week 3
A00 5 13 16
A01 70 120 112
Figure 14. Weekly series graph with polio on “special alert”
Table 9. Polio (A80) on Special Alert
Disease
Actual Third
Quartile Value
Baseline Value Maximum Value
A76 12 15 13
A80 0 0 1
J60 20 30 21
A98 - 4 0 0 0
Figure 15. Sample demographic analysis
5. CONCLUSION
This proposed web-based epidemic alert system is a first-line step to standardizing web-based
disease surveillance systems. It is a grassroots model that leverages on simplicity and “traditionality” as

3827
against complexity and “sophistication” to give timely epidemic alerts that can be disseminated to various
stakeholders for appropriate action.
7. FURTHER WORK
A model for grading the severity of an epidemic and the corresponding response for each stage
would be critically considered in the future. A more comprehensive and effective algorithm for epidemic
detection would also be developed.
ACKNOWLEDGEMENT
This paper was partly sponsored by Covenant University, Ota, Ogun State, Nigeria.
REFERENCES
[1] J. Azeta, et al., “A Plan for Igniting Nigeria‟s Industrial Revolution”, International Journal of Scientific &
Engineering Research, vol. 7, p. 489, 2016.
[2] K. O. Okokpujie, et al., “An Intelligent Online Diagnostic System With Epidemic Alert”, An Intelligent Online
Diagnostic System With Epidemic Alert, vol. 2, 2017.
[3] D. M. Morens, et al., "What is a Pandemic?," ed: The University of Chicago Press, 2009.
[4] E. Christaki, “New Technologies in Predicting, Preventing and Controlling Emerging Infectious Diseases”,
Virulence, vol. 6, pp. 558-565, 2015.
[5] N. Collier, “Uncovering Text Mining: A Survey of Current Work on Web-based Epidemic Intelligence”, Global
Public Health, vol. 7, p. 731Á749, 2012.
[6] J. L. Gardy and N. J. Loman, “Towards a Genomics-Informed, Real-Time, Global Pathogen Surveillance System”,
Nature Reviews Genetics, vol. 19, p. 9, 2018.
[7] O. Oluwagbemi, et al., “Ebinformatics: Ebola Fuzzy Informatics Systems on the Diagnosis, Prediction and
Recommendation of Appropriate Treatments for Ebola Virus Disease (EVD)”, Informatics in Medicine Unlocked,
vol. 2, pp. 12-37, 2016.
[8] E. Adetiba, et al., “Experimental Investigation of Frequency Chaos Game Representation for in Silico and Accurate
Classification of Viral Pathogens from Genomic Sequences”, in International Conference on Bioinformatics and
Biomedical Engineering, pp. 155-164, 2017.
[9] S. Yan, et al., “Utility and Potential of Rapid Epidemic Intelligence from Internet-based Sources”, International
Journal of Infectious Diseases, vol. 63, pp. 77-87, 2017.
[10] S. Pollett, et al., “Internet-based Biosurveillance Methods for Vector-Borne Diseases: Are they Novel Public Health
Tools or just Novelties?," PLoS Neglected Tropical Diseases, vol. 11, p. e0005871, 2017.
[11] N. Schellpfeffer, et al., “Web-Based Surveillance of Illness in Childcare Centers”, Health security, vol. 15,
pp. 463-472, 2017.
[12] M. Dion, et al., “Big Data and the Global Public Health Intelligence Network (GPHIN)”, Canada Communicable
Disease Report, vol. 41, p. 209, 2015.
[13] S. H. Olson, et al., “Drivers of Emerging Infectious Disease Events as a Framework for Digital Detection,"
Emerging Infectious Diseases, vol. 21, p. 1285, 2015.
[14] J. Singh and V. Mansotra, “Salt Contamination Calculation in Insulators During Monsoon using Artificial Neural
Network”, Indonesian Journal of Electrical Engineering and Informatics (IJEEI), vol. 5, pp. 304-308, 2017.
[15] E. O. Nsoesie, et al., “New Digital Technologies for the Surveillance of Infectious Diseases at Mass Gathering
Events," Clinical Microbiology and Infection, vol. 21, pp. 134-140, 2015.
[16] M. A. Al-garadi, et al., “Using Online Social Networks to Track a Pandemic: A Systematic Review”, Journal of
Biomedical Informatics, vol. 62, pp. 1-11, 2016.
[17] C. Rhee, et al., “Syndromic Surveillance System for Korea–US Joint Biosurveillance Portal: Design and Lessons
Learned”, Health security, vol. 14, pp. 152-160, 2016.
[18] K. Al-Surimi, et al., “The Potential of Social Media and Internet-based Data in Preventing and Fighting Infectious
Diseases: From Internet to Twitter”, in Emerging and Re-emerging Viral Infections, ed: Springer,
pp. 131-139, 2016.
[19] V. S. Susan and T. Christopher, “Notice of Retraction Anonymising the Sparse Dataset: A New Privacy
Preservation Approach while Predicting Diseases”, Indonesian Journal of Electrical Engineering and Informatics
(IJEEI), vol. 4, pp. 205-218, 2016.
[20] J. Choi, et al., “Web-based Infectious Disease Surveillance Systems and Public Health Perspectives: A Systematic
Review”, BMC Public Health, vol. 16, p. 1238, 2016.
[21] E. Samoff, et al., “Integration of Syndromic Surveillance Data into Public Health Practice at State and Local Levels
in North Carolina”, Public Health Reports, vol. 127, pp. 310-317, 2012.
[22] N. Mirza, et al., “Steps to a Sustainable Public Health Surveillance Enterprise A Commentary from the
International Society for Disease Surveillance”, Online Journal of Public Health Informatics, vol. 5, p. 210, 2013.
[23] M. O. Lwin, , et al., “Lessons From the Implementation of Mo-Buzz, a Mobile Pandemic Surveillance System for
Dengue”, JMIR Public Health and Surveillance, vol. 3, 2017.

 ISSN: 2088-8708
3828
[24] L. Chimuanya and E. Ajiboye, “Socio-Semiotics of Humour in Ebola Awareness Discourse on Facebook”, in
Analyzing Language and Humor in Online Communication, ed: IGI Global, 2016, pp. 252-273.
[25] C. Abat, et al., “Traditional and Syndromic Surveillance of Infectious Diseases and Pathogens”, International
Journal of Infectious Diseases, vol. 48, pp. 22-28, 2016.
[26] F. Ayeni, et al., “Using Big Data Technology to Contain Current and Future Occurrence of Ebola Viral Disease and
other Epidemic Diseases in West Africa”, in International Conference in Swarm Intelligence, pp. 107-114, 2015.
[27] G. Eysenbach, “Infodemiology and Infoveillance: Framework for an Emerging Set Of Public Health Informatics
Methods to Analyze Search, Communication and Publication Behavior on the Internet”, Journal of Medical
Internet Research, vol. 11, 2009.
[28] M. H. Mamoon, et al., “Visualization for Information Retrieval based on Fast Search Technology”, Indonesian
Journal of Electrical Engineering and Informatics (IJEEI), vol. 1, pp. 27-42, 2013.
[29] S. E. Davies and J. R. Youde, “Surveillance, Response, and Responsibilities in the 2005 International Health
Regulations”, The Politics of Surveillance and Response to Disease Outbreaks: The New Frontier for States and
Non-state Actors, p. 9, 2015.
[30] D. Conesa, et al., “Bayesian Hierarchical Poisson Models with a hidden Markov Structure for the Detection of
Influenza Epidemic Outbreaks”, Statistical Methods in Medical Research, vol. 24, pp. 206-223, 2015.
[31] T. Vega, et al., “Influenza Surveillance in Europe: Comparing Intensity Levels Calculated using the Moving
Epidemic Method”, Influenza and Other Respiratory Viruses, vol. 9, pp. 234-246, 2015.
[32] A. A. Al-Qahtani, et al., “Zika Virus: A new Pandemic Threat”, The Journal of Infection in Developing Countries,
vol. 10, pp. 201-207, 2016.
[33] R. A. Ashton, et al., “The Usefulness of School-based Syndromic Surveillance for Detecting Malaria Epidemics:
Experiences from a Pilot Project in Ethiopia”, BMC Public Health, vol. 16, p. 20, 2015.
[34] P. Hitchcock, et al., “Challenges to Global Surveillance and Response to Infectious Disease Outbreaks of
International Importance”, Biosecurity and Bioterrorism: Biodefense Strategy, Practice, and Science, vol. 5,
pp. 206-227, 2007.
[35] S. Liu, et al., “epiDMS: Data Management and Analytics for Decision-Making from Epidemic Spread Simulation
Ensembles”, The Journal of Infectious Diseases, vol. 214, pp. S427-S432, 2016.
[36] R. F. Olanrewaju and A. W. Azman, “Intelligent Cooperative Adaptive Weight Ranking Policy via dynamic aging
based on NB and J48 classifiers”, Indonesian Journal of Electrical Engineering and Informatics (IJEEI), vol. 5,
pp. 357-365, 2017.
[37] P. A. Idowu, et al., Web-Based Surveillance System for Periodontal Disease for Nigeria.
[38] K. E. L. Raul Isea, “Proposal for an Early Warning System Against an AH1N1 Influenza Pandemic in America”,
Mac Octagon Journals, vol. Vol 1, pp. pp. 65-71, 2017.
[39] G. Texier, et al., “Outbreak Definition by Change Point Analysis: A Tool for Public Health Decision?”, BMC
Medical Informatics and Decision Making, vol. 16, p. 33, 2016.
[40] Z. Harchaoui, et al., “Kernel Change-Point Analysis”, in Advances in Neural Information Processing Systems,
pp. 609-616, 2015.
[41] H. D. Teklehaimanot, et al., “Alert Threshold Algorithms and Malaria Epidemic Detection”, Emerging Infectious
Diseases, vol. 10, p. 1220, 2004.
[42] S. N. Chinedu, et al., “Association Between Age, Gender and Body Weight in Educational Institutions in Ota,
Southwest Nigeria”, Asian Journal of epidemiology, vol. 10, pp. 144-149, 2017.

Epidemic Alert System: A Web-based Grassroots Model

More Related Content

What's hot (20)

Similar to Epidemic Alert System: A Web-based Grassroots Model (20)

More from IJECEIAES (20)

Recently uploaded (20)

Epidemic Alert System: A Web-based Grassroots Model