IAES International Journal of Artificial Intelligence (IJ-AI)

Vol. 13, No. 4, December 2024, pp. 4062~4070

ISSN: 2252-8938, DOI: 10.11591/ijai.v13.i4.pp4062-4070  4062

Implementation of fuzzy logic approach for thalassemia

screening in children

Erliyan Redy Susanto1,2, Admi Syarif1, Warsito1, Khairun Nisa Berawi3,

Putu Ristyaning Ayu Sangging3, Agus Wantoro2
1
Faculty of Mathematics and Natural Science, Lampung University, Bandar Lampung, Indonesia
2
Department of Information System, Faculty of Engineering and Computer Science, Universitas Teknokrat Indonesia,
Bandar Lampung, Indonesia
3
Department of Biomolecular, Biochemistry and Physiology, Faculty of Medical, Lampung University, Bandar Lampung, Indonesia

Article Info ABSTRACT

Article history: Thalassemia is one of the most dangerous blood disorders that can lead to
severe complications. It is an inherited disease, usually detected after a child
Received May 22, 2023 is two to four years old. Identification of thalassemia is a complex task,
Revised Nov 19, 2023 involving many variables. Doctors generally diagnose thalassemia by using a
Accepted Dec 6, 2023 complete blood count (CBC) and high-performance liquid chromatography
(HPLC) test results. However, HPLC tests are expensive and time-
consuming, hence the need for other methods to identify thalassemia. There
Keywords: are many studies on the application of artificial intelligence for medical
applications. In this study, we developed a new fuzzy-based approach to
Artificial intelligence identify thalassemia based on a patient’s blood laboratory results. First, we
Blood disorder analyzed the CBC data for blood disorder prediction. Secondly, we adopt the
Fuzzy approach test results of peripheral blood smear (PBS) to identify whether the person
Inherited disease has thalassemia. We conducted several experiments using 30 (thirty) hospital
Thalassemia patient data and the results were compared with the results provided by
experts. The experimental results show that the system can determine blood
disorders with 93% accuracy and 100% precision in thalassemia prediction.
This system is very effective to help doctors in diagnosing thalassemia
patients.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Admi Syarif
Department of Computer Science, Faculty of Mathematics and Natural Sciences, Lampung University
St. Prof. Dr. Sumantri Brojonegoro No.1, Bandar Lampung, Lampung 35145, Indonesia
Email: [email protected]

1. INTRODUCTION
Blood disorders are known to require very expensive medical expenses. Some blood disorders, such
as thalassemia, hemophilia, blood clots, blood cancer, leukemia, lymphoma, and myeloma, require a lot of
money. Diagnosing thalassemia is a challenging task. In addition to cancer, it is considered one of the most
dangerous diseases in the world [1]. Thalassemia is one type of blood disorder due to genetic characteristics
derived from both parents [2], [3]. Another possibility that it can cause is when a gene mutation occurs [4].
This disease is commonly known from infancy [5], [6]. In some cases, thalassemia is discovered after a
person grows up [4]. When treatment is given late, the disease can cause growth delays in the child and other
complications. Thalassemia is also divided into two categories: transfusion-dependent and nontransfusion-
dependent [7]. The easiest way to perform a blood transfusion is by knowing the hemoglobin levels in the
blood [7], [8]. If it is below the standard value, a blood transfusion should be initiated immediately.
Thalassemia patients often experience anemia which can cause fatigue, weakness, and difficulty in

Journal homepage: http://ijai.iaescore.com

Int J Artif Intell ISSN: 2252-8938  4063

performing physical activity. They may also experience delayed growth and abnormal bone development.
Most of these cases are found in children over the age of two years. Thalassemia detection can be done
through a blood test in a laboratory, commonly called screening. Thalassemia screening has several
components of blood tests, including complete blood count (CBC) and high-performance liquid
chromatography (HPLC) [9], [10]. Doctors usually use CBC to predict a blood disorder [11], [12]. Some of
the CBC components that have been used include hemoglobin (HGB), mean corpuscular volume (MCV), and
mean corpuscular hemoglobin (MCH) [13]. We found the problem that HPLC inspection is expensive, and
the consumption time is prolonged [10], [14], [15]. Thus, our study aims to use a new fuzzy-based approach
to minimize costs and accelerate the screening time for thalassemia.
Artificial intelligence (AI) is one of the fields of computer science that is constantly developing and
widely applied in various fields [16]. In multiple studies, AI has been widely used to solve problems related
to medicine, agriculture, and business [17], [18]. It is also used to predict the presence of blood disorders
[18], [19]. A significant problem in AI applications is dealing with non-deterministic data or information.
One popular approach to this issue is to adopt a fuzzy approach. Since Zadeh [20] introduced it, fuzzy has
been widely used for various applications to solve various problems of disease in humans [21], animals [22],
and also plants [23]. A fuzzy approach to predicting thalassemia in children can overcome several challenges
[24]. Fuzzy logic can help analyze interpretations that are incomplete, ambiguous, or subject to interpretation.
This problem makes them suitable for dealing with thalassemia’s complex and diverse clinical picture. The
proposed method uses a fuzzy approach to provide a more accurate and efficient method for predicting
thalassemia in children.
Thalassemia is very important to study and many studies on the use of AI for thalassemia problems
have been conducted. The use of AI has addressed a number of challenges associated with thalassemia. We
have summarized some AI approaches to solve some problems in thalassemia presented in Table 1.

Table 1. AI research on thalassemia

Year Methods Topics Authors
2010 Multi-layer perceptron, k-nearest neighbors, The effeciency of data types for classification Paokanta et al. [25]
bayesian networks, naïve-Bayes, and performance of machine learning techniques for
multinomial logistic regression screening β-Thalassemia
2016 Fuzzy logic Design of a fuzzy model for thalassemia disease Thakur et al. [11]
diagnosis: using mamdani type fuzzy inference
system (FIS)
2017 Fuzzy logic Thalassemia risk prediction model using fuzzy Thakur and Raw [12]
inference systems: an application of fuzzy logic
2017 Fuzzy logic Using fuzzy logic for improving daily clinical Santini et al. [23]
care of β-Thalassemia patients
2018 Support vector machine Detection of β thalassemia carriers by red cell Roth et al. [13]
parameters obtained from automatic counters
using mathematical formulas
2019 K-nearest neighbor and naïve Bayes Classification of thalassemia data using k-nearest Siswantining et al. [24]
neighbor and naïve Bayes
2020 Random forest Classification of thalassemia data using random Aszhari et al. [17]
forest algorithm
2020 Machine learning and deep learning Artificial intelligence in hematology: current Radakovich et al. [18]
challenges and opportunities
2021 Ensemble classifiers Classification of β-thalassemia carriers from red Sadiq et al. [16]
blood cell indices using ensemble classifier
2022 Machine learning and deep learning A review of artificial intelligence applications in Alaoui et al. [19]
hematology management: current practices and
future prospects

Table 1 shows that various thalassemia problems worked a lot with AI. Some studies on AI with
fuzzy logical approaches have excellent prediction accuracy [23], [26], [27]. Therefore, we use this approach
to solve problems in thalassemia. This article aims to predict blood disorder in a child, focusing on
thalassemia with a fuzzy approach. Our systems require data from laboratory tests such as CBC and
peripheral blood smear (PBS). Thalassemia, referred to in this article, is a beta (β) type. This type of
Thalassemia is most common in Indonesia and other Asian countries. This article perfected previous research
on classifying Thalassemia [11], [28] by adding PBS as a new parameter.
This article consists of two stages. First, the system detects blood disorders by analyzing CBC
laboratory results containing HGB, MCV, and MCH. Second, the CBC analysis results that detect blood
abnormalities are then matched with expert knowledge of thalassemia symptoms on PBS. Our system will
provide a prediction of the patient’s condition based on the analysis performed using CBC and PBS data.

Implementation of fuzzy logic approach for thalassemia screening in children (Erliyan Redy Susanto)
4064  ISSN: 2252-8938

Thus, the system can recommend the patient’s diagnosis result to the doctor. In this case, we use fuzzy
method to make thalassemia prediction using CBC and PBS data as an alternative to CBC and HPLC data.
We propose a new approach of using CBC and PBS to predict thalassemia as an alternative to CBC and
HPLC which are commonly used by doctors today. We evaluate it by comparing the performance of our
developed system with the opinion of an expert (clinical pathologist) and assess it is potential. We have
conducted a study using laboratory data with and without thalassemia to develop and evaluate a fuzzy
prediction method for thalassemia. Thalassemia screening using fuzzy approach with CBC and PBS data
input is our contribution in this study. The study results show that our model has a very good performance in
identifying thalassemia.

2. METHOD
2.1. Frameworks
The classification of diseases with a fuzzy approach is designed according to a predetermined
framework. The framework is a working conceptual description of the thalassemia classification system. The
framework that we have developed is presented in Figure 1. The main parts of Figure 1 are the knowledge
base, inference engine, and user interface. The knowledge base is built on the clinical pathologist’s
knowledge of blood disorders and thalassemia. The inference engine draws conclusions based on the
information stored in the knowledge base. The inference engine is responsible for reasoning about the
knowledge base and using it to solve problems or answer questions. The inference engine follows a set of
rules or logical steps programmed into the system. The user interface is in the form of a program display that
is designed according to its function. Doctors use the user interface to provide CBC and PBS input.
Furthermore, the doctor will get information on predictions of blood disorders and Thalassemia according to
the knowledge of a clinical pathologist.

CBC

User interface

Blood disorder
prediction
Knowledge Inference
base engine PBS
clinical pathologist Thalassemia Doctor
prediction

Figure 1. Framework of thalassemia screening

2.2. Prediction of blood disorders

Doctors usually perform an anamnesis, a physical examination of the patient, and are supported by
the results of blood tests, i.e., CBC. This section uses a fuzzy approach to determine the suspected blood
disorder. The fuzzy approach is used because the opinion of doctors about the standard threshold values on
the CBC results has a tolerance. Based on previous literature [11], [16], CBC data: HGB, MCV, and MCH.
Based on our interviews, clinical pathologists believe that HGB is the first indicator used to suspect the
presence of blood disorders, then to assess other indicators.

2.3. Thalassemia prediction

The doctor’s diagnosis leads to a blood disorder or not blood disorder. If the diagnosis leads to
blood disorders, proceed with a more detailed blood test, namely PBS. Based on the results of the PBS, it will
then be compared with the characteristics of thalassemia disease. In general, the characteristics of thalassemia
based on PBS include microcytic, polychromasia, hypochromic, teardrop, and target cells [7], [29], [30]. The
thalassemia prediction section will provide output in the form of patient information indicated by thalassemia
or not thalassemia.

Int J Artif Intell, Vol. 13, No. 4, December 2024: 4062-4070

Int J Artif Intell ISSN: 2252-8938  4065

2.4. Knowledge based

This study uses CBC, including HGB, MCV, and MCH. The suspected presence of blood disorders
such as thalassemia can be known based on these values [31]. Table 2 is a knowledge base on the prediction
of blood disorders based on CBC.

Table 2. Knowledge-based for blood disorders based on CBC

No Parameters References Disorders units
1 HGB 10.8-12.8 <10.8 or >12.8 g/dL
2 MCV 73-101 <73 or >101 fL
3 MCH 23-31 <23 or >31 Pg

CBC values that do not match the reference values indicate that patients are strongly suspected of
having blood disorders. Doctors recommend that the patient undergo further examination, namely PBS, to
ensure that there are suspected blood disorders. Based on PBS [29], [30], [32], thalassemia has several unique
characteristics: microcytic, hypochromic, teardrop, polychromasia, and target cells. The opinion of expert
states that the patient will be called a carrier if he is microcytic and hypochromic. At the same time, it is
called thalassemia if it has other signs, such as teardrops, polychromasia, and target cells.

2.5. Fuzzy membership function

Based on the knowledge base on blood disorders, determine the parameters used and create a
mathematical function in a fuzzy membership function [33]. The existence of expert tolerance for the
reference value of each parameter used is why we use a fuzzy approach. We construct fuzzy membership
functions and curves for each CBC parameter [11] based on laboratory data on blood disorders presented in
Table 3.

Table 3. Membership function and curves

Parameter Membership function Curves
HGB 1 𝑥 ≤ 10
10.8 − 𝑥
𝑥 [7.2] = { 10 ≤ 𝑥 ≤ 10.8
10.8 − 10
0 𝑥 > 10.8

MCV 1 𝑥 ≤ 71
73 − 𝑥
𝑥 [72] = { 71 ≤ 𝑥 ≤ 73
73 − 71
0 𝑥 > 73

MCH 1 𝑥 ≤ 21
23 − 𝑥
𝑥 [22.2] = { 21 ≤ 𝑥 ≤ 23
23 − 21
0 𝑥 > 23

For example, we consider patient x with CBC data with HGB values of 7.20, MCV of 72.00, and
MCH of 22.20. Based on the membership function in Table 2, the fuzzy values for blood disorders are HGB:
1.00, MCV: 1.00, and MCH: 0.40. As for not blood disorders, the values are HGB: 0.00, MCV: 0.00, and
MCH:0.60. Then we create fuzzy weights for 40% HGB, 30% MCV, and 30% MCH. So, we get a fuzzy
weight for blood disorders of 0.82 and for not blood disorders of 0.18. Based on the fuzzy weights, it can be
concluded that there is a suspicion that the patient has a blood disorder.
Furthermore, a PBS-based examination was carried out on patients suspected of having blood
disorders. We put a “1” if the mark is found and a “0” if it is not found. Signs in question include microcytic,
hypochromic, teardrop, polychromasia, and target cells [30]. The PBS data of patient x showed microcytic: 1,
hypochromic: 1, teardrop: 1, polychromasia: 1, and target cells: 1. Based on the literature [29] and expert
knowledge, PBS is a method commonly used by doctors to diagnose blood disorders. Therefore, the
conclusion of the diagnosis based on the PBS value is that the patient suffers from thalassemia.

Implementation of fuzzy logic approach for thalassemia screening in children (Erliyan Redy Susanto)
4066  ISSN: 2252-8938

3. RESULTS AND DISCUSSION

3.1. Evaluation
The systems we have developed have been tested to evaluate the results of this study. We have as
many as 30 lab data sets of CBC and PBS data. The trial was conducted using patient data obtained from the
hospital. The results are compared to the opinions of experts (clinical pathologist at a public hospital in
Lampung Indonesia). The results of this experiment are presented in Table 4.

Table 4. Test results predictive of blood disorders

No Parameter Expert Proposed system Results
HGB MCV MCH
1 12.90 73.70 25.90 Not Not True
2 5.90 49.30 12.00 Blood disorders Blood disorders True
3 7.20 72.00 22.20 Blood disorders Blood disorders True
4 9.20 98.30 30.80 Blood disorders Blood disorders True
5 13.00 82.30 29,10 Not Not True
6 11.70 59.90 18.80 Not Not True
7 11.80 74.30 27.60 Not Not True
8 8.80 90.50 28.90 Blood disorders Blood disorders True
9 4.50 85.70 29,20 Blood disorders Blood disorders True
10 13.70 79.20 27,60 Blood disorders Blood disorders True
11 11.20 83.50 28.00 Not Not True
12 12.60 78.80 26.80 Not Not True
13 11.30 76.40 25.50 Not Not True
14 15.50 83.60 29.90 Blood disorders Blood disorders True
15 13.80 82.20 28.60 Blood disorders Blood disorders True
16 12.50 77.50 27.60 Not Not True
17 7.30 53.50 14.90 Blood disorders Blood disorders True
18 8.70 72.90 23.30 Blood disorders Blood disorders True
19 9.70 82.10 26,40 Blood disorders Blood disorders True
20 16.70 65.90 21.70 Blood disorders Blood disorders True
21 3.60 91.70 27.30 Blood disorders Blood disorders True
22 11.30 69.00 10:60 Not Not True
23 10.00 92.20 30.10 Blood disorders Not False
24 13.90 84.00 28.90 Blood disorders Blood disorders True
25 4.90 110.10 38.00 Blood disorders Blood disorders True
26 6.70 77.20 25.00 Blood disorders Blood disorders True
27 3.50 80.20 27.80 Blood disorders Blood disorders True
28 5.00 66.70 20.30 Blood disorders Blood disorders True
29 10.80 62.50 19.90 Not Not True
30 14.00 74.00 25.60 Not Blood Disorders False

Table 4 shows that clinical pathologists believe that 8 patients have blood disorders and 11 do not.
The system we developed also predicts the same number. However, there are differences in predictions from
data 23 and 30. The predictions from the two data sets are contradictory to each other. The classification
system for blood disorders that we developed gave results including 18 patients identified as having blood
disorders (true positive), 1 patient with not blood disorder identified as a blood disorder (false positive), 1
patient with blood disorder identified as not blood disorder (false negative), and 10 patients with not blood
disorder were identified as not blood disorder (true negative). We present the confusion matrix in Table 5.

Table 5. Confusion matrix for blood disorders

Negative predictions Positive predictions
Actual negatives 10 1
Actual positives 1 18

Table 5 shows that there were two inaccurate predictions, one wrong prediction for thalassemia and
one wrong prediction for not thalassemia. We can evaluation metrics from these results, such as accuracy,
sensitivity, and specificity. These show that the system has an accuracy value of 93.3%, a sensitivity or recall
value of 94.7%, a specificity value of 90.9%, and a precision value of 94.7%. We illustrate the overall
evaluation metric given by this system in the Figure 2.
We conducted another experiment to evaluate the prediction of thalassemia. The prediction is based
on PBS information about the patient. We have already compared the results with those given by clinical
pathologists. This comparison is presented in Table 6.

Int J Artif Intell, Vol. 13, No. 4, December 2024: 4062-4070

Int J Artif Intell ISSN: 2252-8938  4067

96%
95% 95% 95%
95%

94% 93%
93%
RESULTS

92%
91%
91%

90%

89%

88%
Accuracy Sensitivity or recall Specificity Precision F1-Score

TYPE OF EVALUATION

Figure 2. System evaluation metric for prediction of blood disorders

Table 6. Thalassemia prediction test results

No Parameter Expert Proposed system Results
Microcytic Hypochromic Teardrops Polychromasia Target cells
1 1 1 1 1 1 Thalassemia Thalassemia True
2 1 1 1 1 1 Thalassemia Thalassemia True
3 1 1 0 1 0 Thalassemia Thalassemia True
4 0 0 0 1 0 Not Not True
5 1 1 0 1 0 Thalassemia Thalassemia True
6 0 0 0 0 0 Not Not True
7 0 0 0 0 0 Not Not True
8 0 0 0 0 0 Not Not True
9 1 1 0 0 0 Thalassemia Thalassemia True
10 1 1 1 0 0 Thalassemia Thalassemia True
11 0 0 0 0 0 Not Not True
12 1 1 0 0 0 Thalassemia Thalassemia True
13 0 0 0 0 0 Not Not True
14 0 0 0 0 0 Not Not True
15 0 0 0 0 0 Not Not True
16 0 0 0 0 0 Not Not True
17 1 0 0 0 0 Not Not True
18 0 0 0 0 0 Not Not True
19 1 1 1 1 1 Thalassemia Thalassemia True

Evaluation must be done to measure the performance of the model we have developed. The
evaluation of the thalassemia classification that we propose gives results including 8 patients who are
thalassemia identified as thalassemia (true positive), 0 patients who are not thalassemia identified as
thalassemia (false positive), 0 patients who are thalassemia identified as not thalassemia (false negative), and
11 patients who were not thalassemia were identified as not thalassemia (true negative). We show the
confusion matrix in Table 7.

Table 7. Confusion matrix for prediction of thalassemia

Negative predictions Positive predictions
Actual negatives 11 0
Actual positives 0 8

Based on expert opinions as shown in Table 7, the accuracy of the test results of the system we
developed to identify Thalassemia using PBS is 100%. However, we feel that this research needs to be
developed further by using more datasets and developing models, for example with fuzzy ensembles [34],
[35]. We hope that this work can be continued by using more actual data to test the quality and accuracy of
the system that has been developed.

Implementation of fuzzy logic approach for thalassemia screening in children (Erliyan Redy Susanto)
4068  ISSN: 2252-8938

4. CONCLUSION
This study provides a novel fuzzy-based approach for thalassemia screening. It adopts a fuzzy-based
approach based on CBC and PBS data. The knowledge used in this research is also based on information
from clinical pathologists. The system has been evaluated using the hospital’s actual patient data and
comparing the results with those provided by experts. Based on CBC and PBS data, our system has very high
accuracy in predicting blood disorders (93% accuracy) and thalassemia (100% accuracy). However, more
data sets are needed to test this model. In the future, this model can adapt the fuzzy ensemble to improve
model performance.

ACKNOWLEDGEMENTS
We thank to Universitas Teknokrat Indonesia for supporting this research. We also thank the
Immanuel Hospital Bandar Lampung, Indonesia, especially Mr. Ns. Dwiantoro, for providing the laboratory
data.

REFERENCES
[1] J. I. Johnsen, C. Dyberg, and M. Wickström, “Neuroblastoma—A neural crest derived embryonal malignancy,” Frontiers in
Molecular Neuroscience, vol. 12, pp. 1–11, 2019, doi: 10.3389/fnmol.2019.00009.
[2] S. Perera et al., “Genotype-phenotype association analysis identifies the role of α globin genes in modulating disease severity of β
thalassaemia intermedia in Sri Lanka,” Scientific Reports, vol. 9, no. 1, pp. 1–9, 2019, doi: 10.1038/s41598-019-46674-y.
[3] M. K. Barker et al., “Serum soluble transferrin receptor concentrations are elevated in congolese children with glucose-6-
phosphate dehydrogenase variants, but not sickle cell variants or α-thalassemia,” The Journal of Nutrition, vol. 147, no. 9, pp.
1785–1794, 2017, doi: 10.3945/jn.117.252635.
[4] A. Tamaddoni, S. K. Astaneh, R. Tabaripour, and H. A. -Niaki, “Krüppel-like factor 1 gene mutations in thalassemia patients
from north iran: report of a new mutation associated with β-thalassemia intermedia,” Hemoglobin, vol. 43, no. 1, pp. 12–17, 2019,
doi: 10.1080/03630269.2019.1567528.
[5] I. K. Kwaifa, M. I. Lai, and S. M. Noor, “Non-deletional alpha thalassaemia: A review,” Orphanet Journal of Rare Diseases, vol.
15, no. 1, pp. 1–12, 2020, doi: 10.1186/s13023-020-01429-1.
[6] G. Kakourou et al., “The clinical utility of PGD with HLA matching: A collaborative multi-centre ESHRE study,” Human
Reproduction, vol. 33, no. 3, pp. 520–530, 2018, doi: 10.1093/humrep/dex384.
[7] V. Viprakasit and S. Ekwattanakit, “Clinical classification, screening and diagnosis for thalassemia,” Hematology/Oncology
Clinics of North America, vol. 32, no. 2, pp. 193–211, 2018, doi: 10.1016/j.hoc.2017.11.006.
[8] K. Tari, P. V. Ardalan, M. Abbaszadehdibavar, A. Atashi, A. Jalili, and M. Gheidishahran, “Thalassemia an update: molecular
basis, clinical features and treatment,” International Journal of Biomedicine and Public Health, vol. 1, no. 1, pp. 48–58, 2018,
doi: 10.22631/ijbmph.2018.56102.
[9] I. Mustafa et al., “Genetic epidemiology of beta-thalassemia in the Maldives: 23 years of a beta-thalassemia screening program,”
Gene, vol. 741, 2020, doi: 10.1016/j.gene.2020.144544.
[10] R. Das et al., “A decision support scheme for beta thalassemia and HbE carrier screening,” Journal of Advanced Research, vol.
24, pp. 183–190, Jul. 2020, doi: 10.1016/j.jare.2020.04.005.
[11] S. Thakur, S. N. Raw, and R. Sharma, “Design of a fuzzy model for thalassemia disease diagnosis: Using mamdani type fuzzy
inference system (FIS),” International Journal of Pharmacy and Pharmaceutical Sciences, vol. 8, no. 4, pp. 356–361, 2016.
[12] S. Thakur and S. N. Raw, “Thalassemia risk prediction model using fuzzy inference systems : an application of fuzzy logic,”
Research Journal of Mathematical and Statistical Sciences, vol. 5, no. 7, pp. 1–8, 2017.
[13] I. L. Roth, B. Lachover, G. Koren, C. Levin, L. Zalman, and A. Koren, “Detection of β-thalassemia carriers by red cell parameters
obtained from automatic counters using mathematical formulas,” Mediterranean Journal of Hematology and Infectious Diseases,
vol. 10, no. 1, pp. 1–10, 2018, doi: 10.4084/MJHID.2018.008.
[14] S. L. Mendiratta, S. Bajaj, S. Popli, and S. Singh, “Screening of women in the antenatal period for thalassemia carrier status:
comparison of NESTROFT, red cell indices, and HPLC analysis,” Journal of Fetal Medicine, vol. 2, no. 1, pp. 21–25, 2015, doi:
10.1007/s40556-015-0036-0.
[15] T. Munkongdee, P. Chen, P. Winichagoon, S. Fucharoen, and K. Paiboonsukwong, “Update in laboratory diagnosis of
thalassemia,” Frontiers in Molecular Biosciences, vol. 7, pp. 1–12, 2020, doi: 10.3389/fmolb.2020.00074.
[16] S. Sadiq et al., “Classification of β-thalassemia carriers from red blood cell indices using ensemble classifier,” IEEE Access, vol.
9, pp. 45528–45538, 2021, doi: 10.1109/ACCESS.2021.3066782.
[17] F. R. Aszhari, Z. Rustam, F. Subroto, and A. S. Semendawai, “Classification of thalassemia data using random forest algorithm,”
Journal of Physics: Conference Series, vol. 1490, no. 1, Mar. 2020, doi: 10.1088/1742-6596/1490/1/012050.
[18] N. Radakovich, M. Nagy, and A. Nazha, “Artificial intelligence in hematology : current challenges and opportunities,” Current
Hematologic Malignancy Reports, vol. 15, no. 3, pp. 203–210, 2020, doi: 10.1007/s11899-020-00575-4.
[19] Y. E. Alaoui, A. Elomri, M. Qaraqe, R. Padmanabhan, and R. Y. Taha, “A review of artificial intelligence applications in
hematology management : current practices and future prospects,” Journal of Medical Internet Research, vol. 24, no. 7, pp. 1–24,
2022, doi: 10.2196/36490.
[20] L. A. Zadeh, “The role of fuzzy logic in the management of uncertainty in expert systems,” Fuzzy Sets Systems, vol. 11, no. 1–3,
pp. 197–198, 1983, doi: 10.1016/S0165-0114(83)80081-5.
[21] P. Lialiou, D. Zikos, and J. Mantas, “Development and evaluation of an expert system for the diagnosis of child autism,” in
Studies in Health Technology and Informatics, 2012, pp. 1185–1187, doi: 10.3233/978-1-61499-101-4-1185.
[22] N. C. E. -Chukwu, “Applications of artificial intelligence in agriculture: a review,” Engineering, Technology and Applied Science
Research, vol. 9, no. 4, pp. 4377–4383, 2019, doi: 10.48084/etasr.2756.
[23] S. Santini et al., “Using fuzzy logic for improving clinical daily-care of β-thalassemia patients,” in 2017 IEEE International
Conference on Fuzzy Systems (FUZZ-IEEE), IEEE, Jul. 2017, pp. 1–6, doi: 10.1109/FUZZ-IEEE.2017.8015545.

Int J Artif Intell, Vol. 13, No. 4, December 2024: 4062-4070

Int J Artif Intell ISSN: 2252-8938  4069

[24] T. Siswantining, A. Kamalia, Z. Rustam, F. Subroto, and A. S. Semendawai, “Classification of thalassemia data using K-nearest
neighbor and naïve Bayes,” International Journal of Advanced Science and Technology, vol. 28, no. 8, pp. 15–19, 2019.
[25] P. Paokanta, M. Ceccarelli, and S. Srichairatanakool, “The effeciency of data types for classification performance of machine
learning techniques for screening thalassemia,” in 2010 3rd International Symposium on Applied Sciences in Biomedical and
Communication Technologies (ISABEL 2010), IEEE, Nov. 2010, pp. 1–4, doi: 10.1109/ISABEL.2010.5702769.
[26] A. Wantoro, A. Syarif, K. Muludi, and K. Nisa, “Implementation of fuzzy-profile matching in determining drug suitability for
hypertensive patients,” IOP Conference Series: Materials Science and Engineering, vol. 857, no. 1, 2020, doi: 10.1088/1757-
899X/857/1/012027.
[27] M. Seera and C. P. Lim, “A hybrid intelligent system for medical data classification,” Expert Systems with Applications, vol. 41,
no. 5, pp. 2239–2249, Apr. 2014, doi: 10.1016/j.eswa.2013.09.022.
[28] E. R. Susanto, A. Syarif, K. Muludi, R. R. W. Perdani, and A. Wantoro, “Implementation of Fuzzy-based Model for Prediction of
Thalassemia Diseases,” Journal of Physics: Conference Series, vol. 1751, no. 1, Jan. 2021, doi: 10.1088/1742-
6596/1751/1/012034.
[29] V. Villatoro and M. To, A laboratory guide to clinical hematology. Edmonton: University of Alberta Library, 2019, doi:
10.29173/oer1.
[30] F. Wang et al., “Differential diagnosis of thalassemia and iron deficiency anemia using the CellaVision advanced red blood cell
software,” International Journal of Laboratory Hematology, vol. 43, no. 4, pp. 853–858, 2021, doi: 10.1111/ijlh.13424.
[31] R. Risoluti, S. Materazzi, F. Sorrentino, C. Bozzi, and P. Caprari, “Update on thalassemia diagnosis: New insights and methods,”
Talanta, vol. 183, pp. 216–222, 2018, doi: 10.1016/j.talanta.2018.02.071.
[32] S. Ali et al., “Current status of beta-thalassemia and its treatment strategies,” Molecular Genetics and Genomic Medicine, vol. 9,
no. 12, pp. 1–14, 2021, doi: 10.1002/mgg3.1788.
[33] G. Selvachandran et al., “A new design of mamdani complex fuzzy inference system for multiattribute decision making
problems,” IEEE Transactions on Fuzzy Systems, vol. 29, no. 4, pp. 716–730, 2021, doi: 10.1109/TFUZZ.2019.2961350.
[34] C. Mahanty, R. Kumar, P. G. Asteris, and A. H. Gandomi, “COVID-19 patient detection based on fusion of transfer learning and
fuzzy ensemble models using CXR images,” Applied Sciences, vol. 11, no. 23, Dec. 2021, doi: 10.3390/app112311423.
[35] C. Mahanty, R. Kumar, and S. G. K. Patro, “Internet of medical things-based COVID-19 detection in CT images fused with fuzzy
ensemble and transfer learning models,” New Generation Computing, vol. 40, no. 4, pp. 1125–1141, Dec. 2022, doi:
10.1007/s00354-022-00176-0.

BIOGRAPHY OF AUTHORS

Erliyan Redy Susanto Doctoral student at the Faculty of Mathematics and

Natural Sciences, Lampung University. Currently he is a lecturer at the Faculty of Engineering
and Computer Science, Universitas Teknokrat Indonesia. He has his Bachelor of Computer
Science, Ahmad Dahlan University, Yogyakarta in 2006. He has his Master degree in
Computer Science from Bogor Agriculture University in 2016. His research interests are
primarily in the field of artificial intelligence for medicine. He can be contacted at email:
[email protected].

Admi Syarif Professor at the Dept. of Computer Science, Faculty of Mathematics

and Sciences, Lampung University, Indonesia. He received a Bachelor of Science in
Mathematics from Padjadjaran University, Indonesia, in 1990 and his Ph.D. in Industrial and
Information System Engineering from Ashikaga Institute of Technology, Japan, in 2004. He
was a director of the research center of Lampung University from 2010 to 2016; and has been
a national research reviewer of the Ministry of Education and Culture, Republic Indonesia,
since 2014. He can be contacted at email: [email protected].

Warsito Professor at the Department of Physics, Faculty of Mathematics and

Natural Science, Lampung University. He completed his bachelor’s degree at the Department
of Physics, Brawijaya University, Malang in 1995. Meanwhile, his master’s degree was
completed in 2000 at Universite De Caen, France. For doctoral education he completed in
2004 at the same university in France. His other experience is as the Education and Culture
Attache at the Indonesian Embassy in Paris. Currently he serves as Deputy for Coordinating
Education Quality Improvement and Religious Moderation. He can be contacted at email:
[email protected]

Implementation of fuzzy logic approach for thalassemia screening in children (Erliyan Redy Susanto)
4070  ISSN: 2252-8938

Khairun Nisa Berawi is a professor candidate from the Faculty of Medicine,

University of Lampung. He earned his medical degree from Sriwijaya University, Palembang
in 1999. In 2008, he completed his master’s degree in public health at Padjadjaran University,
Bandung. She successfully completed her doctoral program in 2018 at Andalas University,
Padang. She is a Sports Physiology Expert in Indonesia with an AIFO degree. She can be
contacted at email: [email protected].

Putu Ristyaning Ayu Sangging is a lecturer at the Faculty of Medicine,

University of Lampung with expertise in Clinical Pathology. She completed medical education
in 2002 at Hang Tuah University, Surabaya. Then she successfully completed her master’s
degree in health and clinical pathology specialist in 2012 at Hasanuddin University, Makassar.
Furthermore, he completed his sub-specialist education at Diponegoro University, Semarang.
She can be contacted at email: [email protected].

Agus Wantoro is a Faculty of Engineering and Computer Science lecturer at the

Universitas Teknokrat Indonesia. He obtained a bachelor’s degree in computer from the
College of Informatics Management and Computer Teknokrat Lampung in 2009. He
completed a master’s degree in computer in 2015 from Budi Luhur University, Jakarta. Then
in 2022 complete his doctoral education at the University of Lampung. The field currently
occupied is artificial intelligence in the medical field. He can be contacted at email:
[email protected].

Int J Artif Intell, Vol. 13, No. 4, December 2024: 4062-4070