Data Mining from Knowledge Cases of COVID-19
MARIYA EVTIMOVA-GARDAIR
University Paris 1 Panthéon- Sorbonne, INRIA,
Paris,
FRANCE
Abstract: - A lot of articles were produced during the pandemic of COVID-19 and continue to be produced.
The article proposes a system for diagnosis of COVID-19 disease. Also nowadays, the presentation of
knowledge and the research for the reasoning algorithms are progressively improving in the domain of
Artificial Intelligence. Besides these, distributed reasoning as a part of data mining has become a solution for
the increasing everyday data amount. As a result, the paper proposes a case-based non-monotonic reasoner for
uncertain and vague COVID-19 information that is appropriate for work with Big Data. Also, a COVID-19
knowledge base model is proposed. The reasoner implements rules for the distribution of the information that
gives the possibility to work with Big data. The proposed reasoning algorithm is applied for COVID-19. It
shows the implementation of the reasoner into the data mining system and the returned results from the system
are evaluated. The results show that the system returns relatively high results concerning the other system
for recommendation.
Key-Words: - COVID-19, rule-based reasoning, case-based reasoning, data mining, reasoning, non-monotonic
reasoning, jColibri.
Received: May 11, 2023. Revised: December 12, 2023. Accepted: January 9, 2023. Published: February 20, 2024.
1 Introduction
In the effort to respond to the overwhelming need
defined by the pandemic of COVID-19 are created
many computer-aided applications for diagnosis and
recommendation. The recent literature that describes
COVID-19 solutions is defined also as the issues
that are an object of continuous improvement such
as methods for analysis and the poor quality of the
data sets, [1]. The reasoning methodologies in
intelligent systems can be divided into rule-based
reasoning and case- based reasoning. Using rules
can define general knowledge and using cases reuse
already-defined knowledge in specific situations.
Each research technique has its advantages and
disadvantages, which have been proven to be mostly
complementary. Combining two or more different
reasoners and hybrid knowledge representation is a
very active research area in artificial intelligence,
[2]. The aim is to create a combined formalism that
uses the advantages of the methods.
The effectiveness of the different hybrid or
integrated approaches has been demonstrated in
several areas of application. The effectiveness of
these studies is because the rules are complementary
in a field of application or solving a problem.
Rule-based systems solve problems from the
start while case-based systems use previously stored
situations to deal with a similar case. Therefore, the
combination of the two approaches is natural and
useful. Complementary advantages and
disadvantages of both intelligent methods enhance
the benefits of their combination. Furthermore, the
combination of fuzzy and probability theory proves
their possibility when using uncertain and vague
information from the query. Those methods are
integrated to improve the quality of the searched
information in Big Data which is particularly
important when performing diagnosis for COVID-
19. Conversely, this work focuses on the
distributed non-monotonic reasoning
methodology. It could be defined the following
technical contribution to the study:
1) Definition and creation of COVID-19 case-
based ontology for feature extraction and
mapping from suspected cases of COVID-19.
2) Proposal of a novel mathematical model for
semantic-based and feature-based case
similarity computation.
3) Incorporation of the proposed reasoning model
into an improved CBR (Case-Based Reasoning)
framework.
4) Implementation of CBR framework which
allows for the detection or classification of
suspected cases of COVID-19 as either positive
or negative.
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
99
Volume 21, 2024
The content of the article is organized into
seven sections. The first section performs the
introduction of the article. The second section
describes the related work concerning reasoning
over COVID-19. The third section represents the
knowledge base model that is used for the reasoner.
The fourth section of the article represents the
reasoning model. The fifth section describes the
proposed COVID-19 data mining system and the
implementation of the proposed Big data reasoning
algorithm. Section 6 proposes the evaluation of the
system. In the end, the study concluded with section
7.
2 Related work
The research activity in reasoning and explanation
as a part of artificial intelligence has been started for
several decades. Moreover, using the classic variant
of propositional logic seems to be not an appropriate
choice when doing reasoning tasks in real
applications because of the presence of conflicting
and inconsistent information. The explanation of
this is with the monotonicity property that uses
consequence relation when implementing formula
into the theory but as a result, is not reaching a
reduction of the consequence set.
Generally, monotonicity relies on the fact that
learning a new set of knowledge is not able to
reduce the already known set of knowledge. Also, it
can be defined that when applying a standard
monotonic reasoning a conclusion that respect
package of premises is still valid if another package
premises is added, [3]. But that property cannot
permit the removal of well-known knowledge as a
contradiction of what can be observed in human-like
reasoning that can be classified as non-monotonic.
As a definition non-monotonic reasoning represents
the possibility of a conclusion with a set of premises
to be removed to update the existing information.
As a consequence of this, non-monotonic
reasoning can have true premises but not
corresponding conclusions. Regarding monotonic
reasoning can be defined that when premises are
true the conclusions necessarily follow them. The
non-monotonicity property supports that the claim
can specify partial premises but when an exception
arises it can be withdrawn.
Also, including new premises can be associated
with retracting described as non-monotonicity,
rather than implementing new conclusions described
as monotonicity). Since the beginning of the
COVID-19 pandemic, many articles have been
produced to propose a solution to the fight against
COVID-19. Article, [4], describes the
epistemological problem of induction in COVID-19.
Another article, [5], describes the uncertainty when
making decisions during the COVID-19 pandemic.
To analyze the COVID-19 data in Mexico is
implemented a non-monotonic behavior of the real
data, [6]. Paper, [7], proposes an investigation of the
spread of COVID-19 by countries that use non-
monotonous relationships. Reasoning over COVID-
19 ontology is described in [8]. An approach of
fuzzy case-based reasoning that evaluates COVID-
19 patients is described, [9]. Another technique used
for COVID-19 are [10], [11]. Some articles are
more general, [12], which also describe the benefits
of using artificial intelligence for searching COVID-
19 data.
3 Representation of Knowledge Base
Model that was used for the
Reasoner
A brief description of the knowledge base schema
that was used for the system is defined in Figure 1.
In the schema are defined the COVID-19 symptoms,
[13]. The symptoms for COVID-19 defined in [13],
and reused in this research are Cough, Fever,
Anosmia, Pneumonia, Acute respiratory distress
syndrome(ARDS), Organ failure, Dyspnea, Nausea
and vomiting, Headache, Diarrhea, Respiratory tract
infections.
Shortness of breath, Rhinorrhea,
Gastrointestinal symptoms, Chest pain/tightness,
Abdominal pain, Muscle pain, Loss of appetite,
PaO2, SaO, Loss of smell, Heart rate, Systolic
pressure, Diastolic pressure, Fatigue, Septic shock,
Sore throat, pH, Temperature, Pharyngeal pain. The
data is taken from the Bulgarian hospital St. George,
then they are adapted in a case-based format that is
suitable for the platform Colibri.
The proposed case-based knowledge data model
is compatible with CBROnto. Rules as well as
additional classes are not represented in the
knowledge model. As presented in Figure 1 CBR
INDEX includes all the features of the precedent,
CBRCASE includes the individual capabilities of
the precedent, and HAS-COMPONENT the two
parts of the precedent diagnosis and precedent
description. In the precedent-based model,
precedents are represented as exemplar concepts and
their attributes are represented as semantic relations
and properties. The values that link attributes can be
taken as precedents defined within the same domain
knowledge model.
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
100
Volume 21, 2024
Fig. 1: Case-based knowledge base schema
The schema represented in Figure 2 provides the
mechanism for retrieving data from the knowledge
base. To personalize and precise the returned
information from the system, it is defined as a user
profile. The user profile includes the personalized
information from the user that is used during
retrieval with the current symptoms provided in the
system to perform the diagnostic and return
appropriate results. The COVID-19 information was
used from, [13].
Fig. 2: Representation of the user profile for
personalization of the result
The schema provides the mechanism for
retrieving data from the knowledge base. The user
profile gives the possibility for each user of the
system to add their specific information that
improves the proposed results from the system.
4 Representation of Non-Monotonic
Reasoning Model
This reasoner is appropriate to use with the
described knowledge base above. Another
interesting research related to the knowledge base
for drugs for COVID-19 is described, [14]. The
proposed reasoning algorithm uses a case-based
reasoning theory and also the rule-based reasoning
theory. The combination of the two algorithms is
defined to propose a quality reasoning solution for
Big Data. The first phase of the algorithm includes
the implementation of rules-based reasoning. To
meet the Big Data challenge the proposed reaso ner
implements distributed reasoning. To realize this
complexity, several defined rules and data are
distributed among separate peers. Using a selection
of a Big set of data inevitably leads to the presence
of uncertainty, because of the incomplete
information. For example, if it is necessary to give a
diagnosis of disease using a query that contains in
the query text such as fever and chest pain. But that
can be used for all COVID-19 diseases or it can be
considered only the COVID-19 disease and then to
conclude with a certain degree of uncertainty.
This step can increase the matching degree, as
the step is characterized by incompleteness and
uncertainty. Two stages can be defined concerning
this:
1) preparation of the separation of the data
2) definition of data distribution
The basic problem observed in stages 1 and 2 is
the presence of certain events with certain rules that
cannot be concluded if they are not related. The
distribution of the rules and cases is strictly defined.
So that the expressions that are connected must if it
is possible to coexist in the same peer.
The distribution is defined so that every peer is
linked with a set of rules and a set of cases.
Each peer has defined entities:
1) basic rules to the concept that present a class of
strict and conditional restrictions
2) concept of the case that presents a class of cases
3) concept of the peer that presents a class of peers
The basic rules have the form: Rule: (m —
n[md]), means that if n holds then m normally
holds with a matching degree md. The cases have
the form: Case: (c—T[md]), which means that the
case is true with matching degree md, where the
activation weight of the rule represents the matching
degree, i.e. degree of belief. As it is described in
[15], [16], the disjunctive belief reasoning is
implemented. This equivalence serves as a way of
using probabilistic knowledge bases to represent
uncertain and vague concepts. Each concept could
be associated with another concept, indicating that
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
101
Volume 21, 2024
there is a correlation. This is presented with a key
value:
• if m then n with activation weight A and matching
degree md
• m is true with activation weight A and matching
degree md
The role of the key value includes a general
description of the rules and a description of the
cases. The general description
Algorithm 1 Algorithm implementing correlation factor
1: Determination of the number of peers equal to the
number of key values
2: For each concept is needed to connect the concept
with the peer that has the same number as the key
value of the case
3: Finding a conclusion about cases and rules for
each peer of the rules describes rules, which can
only be used when there is a specific case. The
description of the cases describes cases that are
derived from the same result in the same general
rules i.e. “If COVID-19 disease has fever with
factor 1”, then this is true for all COVID-19 type
diseases, as a rule, fever is common. This means
that the procedure for conclusions needs to be
applied only once for each case. Once the key
value is determined for each concept, the correlation
of the two concepts is defined as follows: If c1 and
c2 are the concepts, then they are related when the
matching degree has the same value.
So the first stage in the method is to distribute
the cases and the rules between the peers according
to the value of the matching degree. For example, if
there is a rule of COVID-19 disease that affect
different parts of the body, then the key is the
frequency of occurrence of a disease, and therefore
the COVID-19 cases can be allocated to peers
depending on the category of the symptoms that
could separate the diseases, i.e. abdominal
symptoms, cardiovascular system symptom,
digestive system symptom, head and neck symptom,
musculosceleton system symptom, nervous system
symptom, neurological and physiological symptom,
nutrition metabolism and development system
symptom, reproductive system symptom, respiratory
system and chest symptom, skin and integumentary
tissue symptom, urinary system symptom, hemic
and immune system. Description of the algorithm
with correlation factor, [15], [17], Figure 3 presents
the proposed model for distributed reasoning. The
key value also determines the number of used peers.
The number of used peers is equal to the
number of used key values. After determining each
peer, each peer contains the number of cases and
rules. Then is applied the algorithm for reasoning.
This process runs independently in each peer
without taking into account other peers, [18]. The
problem could occur in situations where two rules
have the same conclusion with different matching
degree values. For example, if there are two rules
COVID-19 diseases have fever with factor 0.5 and
COVID-19 disease always have chest pain, then the
second rule should be applied instead of the first.
Furthermore, these conflicts can arise in situations
where there are exceptional classes (such as heart
disease) and special rules, so extra classes are
treated as opposed to the total class. The algorithm
for the reasoning of each peer is presented below:
Reducing the facts serves as a way to stop running
rules with low priority, so that if a rule that starts
has priority 2, then by reducing the fact that causes
running, the rule with priority 1 is not going to be
implemented. Finally, a solution for the most similar
case is retrieved and proposed to the user.
Fig. 3: Representation of algorithm for distributed reasoning
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
102
Volume 21, 2024
Finally, a solution for the most similar case is
retrieved and proposed to the user.
Algorithm 2 Algorithm that is executed to each peer
1: Implement priority 2 on each exceptional rule
2:Implement priority 1 of the rules without
exceptions
3: Run the rule engine from higher to lower priority
4: Cancellation of the facts when the rule run
for the most similar case is retrieved and proposed
to the user.
This algorithm for reasoning is suitable for
working with Big Data and can handle vague and
uncertain data that include the belief-defined case to
be a conclusion. The schema of the reasoner
construction is shown in Figure 3. The reasoning
algorithm is created using SWRL rules.
5 Description of COVID-19 Data
Mining System with the Proposed
Reasoning Algorithm
The developed application model with the proposed
reasoner is shown in Figure 4 as the application is
divided into three layers and each layer has a
specific task. For the practical realization of the
model is used the Java library jColibri, [19] and
Jess. Furthermore, case-based reasoning is used also
in [20], which is an article related to COVID-19.
The primary layer maintains the communication
of the knowledge graph with cases. Then the
application layer contains the retrieved data from
the cases. This layer is case-based and is very
important for the application. The interface layer
can accept the request from the user application and
return the nearest event. In the application, only the
steps for presentation and searching. The model
does not include an adaptation of the cases and their
storage, which describe clearly the structure of the
proposed reasoner. The reasoner is presented in two
steps. The first step is reasoning with uncertain and
vague information.
Fig. 4: Personalized data mining system in layered
model
6 Evaluation of the Results from the
System
Basically for estimation of the results of the search
system are accepted three parameters Precision,
Recall, and F-Measure. The aim of the semantic
search within COVID-19 knowledge base is to
increase precision and recall, where in Figure 5
present the formulas.
Fig. 5: Formulas for precision, recall and F-measure
For the evaluation of the system, 223 user requests
for COVID-19 were taken from various internet
sources. The data is separated into eleven categories
concerning the symptoms that affect different parts
of the body: Head and neck symptoms,
Musculosceleton system symptom, Nervous system
symptoms, Respiratory system and chest symptom,
Skin and integumentary tissue symptom. Then the
proposed system is evaluated by an expert and the
results returned from the system and the expert are
compared. The proposed system analysis does not
account for the results when both the expert and the
system give different results for a given disease, so
the negative predicted value parameter cannot be
estimated. The system is making general diagnostics
for COVID-19 disease.
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
103
Volume 21, 2024
The system is one shot i.e. returns a single
result. Figure 6 represents the evaluation of the
system concerning the Precision.
Fig. 6: Precision of the system
Figure 7 presents the evaluation of the system
concerning the Recall.
Fig. 7: Recall of the system
Figure 8 shows the F-measure of the system
separated into the categories of the system. It
calculates the accuracy of the system with medical
information about COVID-19 diseases.
Fig. 8: F- measure of the system
The precision is defined as 99% and the recall
as 93% that gives the quality of the system returned
results, [21], [22], [23], [24]. Figure 9 represents the
analysis of different COVID-19 applications defined
in the literature concerning their evaluation results.
The results from the analysis defined in Figure 9
show that the proposed system has relatively high
results in comparison with the other systems. The
system uses Iris prediction cost algorithm for the
calculation of the accuracy.
Fig. 9: Representation of comparative analysis
The accuracy of the system is defined as 95.5%,
[25]. The algorithm is a part of the Colibri
framework that also provides tools for testing.
After evaluation of the system, the results show
that the system has relatively high quality
concerning the other systems.
The analysis methods described in [26], are
used to evaluate the system as suitable for Big Data.
As a conclusion from the analysis of [27], is defined
that speed is crucial for Big Data systems. A
computer with the following system configuration is
used to test the system: Intel Core i9-13980HX, 32
Go RAM. The results are described in Table 1.
Table 1. Comparative results
Using
case-based
algorithm not
adapted to
work with Big
Data
Time to respond
7524ms
The results from Table 1 show that the proposed
algorithm that is adapted to work with Big Data is
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
104
Volume 21, 2024
faster and has a smaller time to respond than the
algorithm not adapted to work with Big Data.
The total time for searching is bigger when
using a standard case-based algorithm that is
available in jColibli than the proposed algorithm
that is adapted for working with Big Data.
Because the standard algorithm proposed by
jColibri does not distribute the data to peers before
searching the information and then indexing the
results. For the creation of the system, the Colibri
framework is available on the internet, [25]. This
software provides easy prototyping functionality
and contains different searching algorithms and
tools for testing. Furthermore, the Colibri
framework is a free case-based framework that
makes it easily accessible and usable.
7 Conclusion
A reasoning algorithm that is proposed combines
case-based and rule-based reasoning. It can be
defined as a case-based non-monotonic reasoner that
is suitable for vague and uncertain information and
can work with Big Data. The proposed non-
monotonic reasoning algorithm uses disjunctive
belief logic that is adapted to work with COVID-19
data. This algorithm is adapted to be case-based and
uses the experience from the previous cases and
improve the results from the system. With the
combination of the cases, the algorithm also uses
rules to provide the possibility to perform
distributed searching and work with Big Data. Also,
it is proposed a case-based knowledge base model
for COVID-19 hospital data. The proposed reasoner
is implemented into a personalized COVID-19
system that gives quality results to the user.
References:
[1] Xingyi Yang, Xuehai He, Jinyu Zhao, Yichen
Zhang, Shanghang Zhang, Pengtao Xie.
Covid-ct-dataset: a ct scan dataset about
COVID-19. arXiv preprint arXiv:2003.13865,
2020,vol. 490, no 10.48550.
[2] Albahri, Ahmed Shihab, Hamid, Rula A.,
Alwan, Jwan K., et al. Role of biological data
mining and machine learning techniques in
detecting and diagnosing the novel
coronavirus (COVID-19): a systematic
review.Journal of Medical Systems, 2020, vol.
44, p. 1-11.
[3] Longo, Luca, Rizzo, Lucas, Dondio,
Pierpaolo. Examining the modeling
capabilities of defeasible argumentation and
non-monotonic fuzzy reasoning. Knowledge-
Based Systems, 2021, vol. 211, p. 106514.
[4] Vega, Carlos. From Hume to Wuhan: an
epistemological journey on the problem of
induction in COVID-19 machine learning
models and its impact upon medical research.
IEEE Access, 2021, vol. 9, p. 97243- 97250.
[5] Stan, Sergiu Octavian, Treapat, Mihai
Laurent, iu, et Draganescu, Corina Elena.
Decisions under uncertainty in the Covid-19
era. Strategica: Preparing for Tomorrow,
Today, p.731-742, 2020.
[6] Aguilar- Madera, Carlos G., Espinosa-
Raredes, Gilberto, Herrera- Hernandez, E. C.,
et al. The spreading of Covid-19 in Mexico: A
diffusional approach. Results in Physics,
2021, vol. 27, p. 104557.
[7] Gonzalez-Val Rafael, Sanz-Gracia, Fernando.
Urbanization and COVID-19 incidence: A
cross-country investigation. Papers in
Regional Science, 2022, vol. 101, no 2, p.
399-415.
[8] Groza, Adrian. Detecting fake news for the
new coronavirus by reasoning on the Covid-
19 ontology. arXiv preprint
arXiv:2004.12330, 2020.
[9] Ahmed Faisal, Hossain Mohammad Shahadat,
Islam Raihan Ul, Karl Andersson. An
evolutionary belief rule-based clinical
decision support system to predict COVID-19
severity under uncertainty. Applied Sciences,
2021, vol. 11, no 13, p. 5810.
[10] Cai, Junzhe, Revesz, Peter Z. A Novel
Spatiotemporal Method for Predicting Covid-
19 Cases. WSEAS Transactions on
Mathematics, 2021, 20: 300-311,
https://doi.org/10.37394/23206.2021.20.31.
[11] Liu, Chenglian; Chen, Sonia CI. Study of
COVID-19 Monitoring System Based on
Block Chain and Anonymity Techniques.
WSEAS Transactions on Mathematics, 2022,
vol.21, p.635-640,
https://doi.org/10.37394/23206.2022.21.74.
[12] Vaishya, Raju, Javaid, Mohd, Khan, Ibrahim
Haleem, et al. Artificial Intelligence (AI)
applications for COVID-19 pandemic.
Diabetes & Metabolic Syndrome: Clinical
Research & Reviews, 2020, vol. 14, no 4, p.
337-339.
[13] Oyelade, Olaide N. et Ezugwu, Absalom E. A
case-based reasoning framework for early
detection and diagnosis of novel coronavirus.
Informatics in Medicine Unlocked, 2020, vol.
20, p. 100395.
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
105
Volume 21, 2024
[14] Khan, Junaed Younus, Khondaker, Md
Tawkat Islam, Hoque, Iram Tazim, et al.
Toward preparing a knowledge base to
explore potential drugs and biomedical
entities related to COVID-19: automated
computational approach. JMIR medical
informatics, 2020, vol. 8, no 11, p. 21648.
[15] Belle, Vaishak. Toward Robots That Reason:
Logic, Probability & Causal Laws. Springer,
2023.
[16] Fu, Yang-Geng, Fang, Geng-Chao, Liu,
Yong-Yu, et al. Disjunctive belief rule-based
reasoning for decision making with
incomplete information. Information Sciences,
2023, vol. 625, p. 49-64.
[17] Hariri, Reihaneh H., Fredericks, Erik M., et
Bowers, Kate M. Uncertainty in big data
analytics: survey, opportunities, and
challenges. Journal of Big Data, 2019, vol. 6,
no 1, p. 1-16.
[18] Insaurralde, Carlos C., Blasch, Erik P., Costa,
Paulo CG, et al. Uncertainty-driven ontology
for decision support system in air transport.
Electronics, 2022, vol. 11, no 3, p. 362.
[19] Alazzam, Malik Bader, Tayyib, Nahla,
Alshawwa, Samar Zuhair, et al. Nursing care
systematization with case-based reasoning and
artificial intelligence. Journal of Healthcare
Engineering, 2022, vol. 2022.
[20] Nurgraheni, Murien, SARI, Irma Permata, et
al. A Case-Based Reasoning for Detection
Coronavirus (Covid-19) Using Cosine
Similarity. In: Conference on Broad Exposure
to Science and Technology 2021 (BEST
2021). Atlantis Press, 2022. p. 178-183.
[21] Gao, Chengcheng, Zhang, Rui, Chen,
Xicheng, et al. Integrating Internet multi
source big data to predict the occurrence and
development of COVID-19 cryptic
transmission. NPJ Digital Medicine, 2022,
vol. 5, no 1, p. 161.
[22] Xu, Xiaowei, Jiang, Xiangao, Ma, Chunlian,
et al. A deep learning system to screen novel
coronavirus disease 2019 pneumonia.
Engineering, 2020, vol. 6, no 10, p. 1122-
1129.
[23] Goodman- Meza, David, Rudas, Akos,
Chiang, Jeffrey N., et al. A machine learning
algorithm to increase COVID-19 inpatient
diagnostic capacity. Plos one, 2020, vol. 15,
no 9, p. e0239474.
[24] Ahmed, Imran, Ahmad, Misbah, Jeon,
Gwanggil, et al. A framework for pandemic
prediction using big data analytics. Big Data
Research, 2021, vol. 25, p. 100190.
[25] Ma, Li, Tan, Tieniu, Wang, Yunhong, et al.
Efficient iris recognition by characterizing
key local variations. IEEE Transactions on
Image processing, 2004, vol. 13, no 6, p. 739-
750.
[26] Novitsky, A. V. The concept and evaluating
of big data quality in the semantic
environment. Problems in Programming,
2022, no 3- 4, p. 260-270.
[27] Recio- Garcia, Juan A., Diaz- Agudo, Belen,
et Gonzalez- Calero, ‘Pedro A. The COLIBRI
platform: tools, features and working
examples. Successful Case-based Reasoning
Applications-2, 2014, p. 55-85.
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The authors equally contributed in the present
research, at all stages from the formulation of the
problem to the final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The authors have no conflicts of interest to declare.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2024.21.10
Mariya Evtimova-Gardair
E-ISSN: 2224-3402
106
Volume 21, 2024