Content uploaded by Sudip Mandal
Author content
All content in this area was uploaded by Sudip Mandal on May 11, 2022
Content may be subject to copyright.
International Journal of Computer Applications (0975 – 8887)
Volume 184 – No.9, April 2022
7
AI based Automated Diagnosis of COVID-19 Patients
Sudip Mandal
ECE Department, Jalpaiguri
Government Engineering College
Jalpaiguri, India
Shankhalika Mallick
ECE Department, Jalpaiguri
Government Engineering College
Jalpaiguri, India
Arkaprava Roy
EE Department, Jalpaiguri
Government Engineering College
Jalpaiguri, India
Arghyadip Paul
EE Department,
Jalpaiguri Government Engineering College
Jalpaiguri, India
ABSTRACT
The recent Corona Virus Disease 2019 (COVID-19)
pandemic has placed severe stress on healthcare systems
worldwide, which is amplified by the critical shortage of
COVID-19 tests. Effective screening of SARS-CoV-2 enables
quick and efficient diagnosis of COVID-19 and can mitigate
the burden on healthcare systems. In this study, a satisfactory
accurate automated diagnosis model of COVID-19 based on
patient symptoms has been proposed by applying several
Artificial Intelligence (AI) models. For training, COVID-19
data has been collected from Israeli Ministry of Health
publicly released data. We have used Artificial Neural
Network and Decision Tree for classification or prediction
purpose. The proposed model predicted COVID-19 test
results with satisfactory accuracy. Hence, this AI based
diagnostic framework can be used, among other
considerations, to prioritize testing for COVID-19 when
testing resources are limited.
Keywords
COVID-19; Artificial Intelligence (AI); Artificial Neural
Network (ANN); Decision Tree; Automated Disease
Diagnosis; Classification
1. INTRODUCTION
The novel corona virus (COVID-19 or SARS-COV-2)
epidemic has been acquainted as a global pandemic. More
than 21.9 crore individuals have been infected worldwide,
leading to more than 45.5 lakhs deaths as of 7th October 2021
[1]. Rapid human-to-human transmission accompanied by
unrevealed nature of the virus had led to this tremendous
outbreak of COVID-19. It affects people in various ways. The
main symptoms are main symptoms are fever, cough, myalgia
or fatigue, sputum, and dyspnea [2]. However, 80% of
patients get well from the disease without facing any serious
complication. It has been observed that one out of every six
affected people get seriously sick and develop difficulty in
breathing due to infection and inflammation in lungs caused
by the Corona virus [3].
This pandemic continues to challenge medical systems
worldwide in many aspects, including sharp increases in
demands for hospital beds and critical shortages in medical
equipment, while many healthcare workers have themselves
been infected. Thus, the capacity for immediate clinical
decisions and effective usage of healthcare resources is
crucial. The most validated diagnosis test for COVID-19,
using reverse transcriptase polymerase chain reaction (RT-
PCR), has long been in shortage in developing countries. This
contributes to increased infection rates and delays critical
preventive measures.
Effective screening enables quick and efficient diagnosis of
COVID-19 and can mitigate the burden on healthcare
systems. Prediction models that combine several features to
estimate the risk of infection have been developed, in the hope
of assisting medical staff worldwide in triaging patients,
especially in the context of limited healthcare resources. To
defeat the COVID-19 outbreak [4], appropriate and evidence-
based actions must be taken worldwide. For this purpose,
prediction models can help not only allocating medical
resources but also raising the preparedness of healthcare
systems involved. In this regard, mathematical, computational
and statistical methods have been utilized to predict if a
person is affected by COVID-19 or not using Artificial
Intelligence (AI) techniques.
In literature, several AI or machine learning models have
already proposed for automated diagnosis of the COVID-19.
Mainly, two different approaches were being used: COVID
affected Lungs X-ray / CT scan images analysis [5], [6] and
symptoms based database [7], [8] analysis using different AI
techniques. In case of X-ray / CT scan images analysis,
different researchers have contributed in this domain using
image processing techniques. Li et al. [9] used CT scan
images to classify the images in either COVID-19 or normal
and they also calculate severity of infection. Hassanien et al.
[10] used support vector machine and multilevel thresholding
to classify the X-ray images of lungs. On the hand, Sing et al.
[11] use multi-objective differential evolution–based
convolutional neural networks to classify CT scan images of
chest affected by COVID-19. Yan et al. [12] used image
processing tools to identify the severity of infection for
COVID-19. Rajinikanth et al. [13] used Harmony Search
(HS) algorithm along with multilevel thresholding to identify
the infection severity in lungs CT scan images that created
due COVID-19. S. Mandal [14] used Elephant Swarm Water
Search Algorithm (ESWSA) for multilevel thresholding based
on Otsu’s and Kapur’s method and further calculation of
severity of infections of lungs images.
On the other hand, several authors also proposed different AI
based models for diagnosis of COVID-19 with the help of
symptoms based clinical data of COVID-19 patients. Khanday
et al. [15] have used different machine learning techniques for
the classification of COVID clinical data. Iwendi et al. [16]
have used boosted random forest algorithm for COVID-19
International Journal of Computer Applications (0975 – 8887)
Volume 184 – No.9, April 2022
8
patient health prediction. Chen et al. [17] used machine
learning techniques for early prediction of mortality rate for
COVID-19. Sudre et al. [18] have developed an app that helps
to predict if a person is affected by COVID-19 or not using by
analyzing the COVID cases and their symptoms. Zoabi et al.
[19] used Gradient boosting predictor for automated diagnosis
of COVID-19 patients using Israeli Ministry of Health data
[28] which is publicly available for the researchers. The
obtained results showed very satisfactory accuracy. Dutta et
al. [20] utilized three different machine learning models
namely bagging algorithm, k-nearest neighbour, and random
forest for prediction. They have used the real-time COVID-19
dataset from India Government website [29] and obtained
satisfactory accuracy. Mei et al. [20] used AI based
techniques for rapid diagnosis of COVID-19 cases to reduce
the burden of health workers.
Through this study, an automated prediction model based on
Artificial Intelligence models are proposed to predict the
COVID-19 through a given set of data comprising of eight
different symptoms. Artificial Neural Network (ANN) and
Binary Decision Tree are employed for classification purpose.
Firstly, a record of confirmed cases of a desired place is taken.
Then, the corresponding data were divided into two parts as
training and test data. The former was used to train the
models, while the latter was used for validation purposes.
Thus, the estimated accumulative confirmed cases of the test
data were compared with those of actual target values. The
rest of the manuscript is structured as follows. Preliminary
background of Artificial Neural Network and Binary Decision
Tree are discussed in next section. Propose methodology and
data collections processes have been elaborated in Section 3.
Results using ANN and Decision tree for classification of
COVID-19 data have been shown in Section 4 followed by
the Conclusion section.
2. THEORETICAL BACKGROUND
In this section, some preliminary concept on Artificial Neural
Network and Decision Tree are discussed those are required
to understand this research work.
2.1 Artificial Neural Network
An ANN [22] is an information processing paradigm that
is inspired by the way the biological nervous system (such as
the brain) that process information. An ANN is configured for
a specific application such as pattern recognition or data
classification [23] through a learning process. An ANN is
typically defined by three types of parameters:
1. The interconnection pattern between the different
layers of neurons. Three types of layers are observed: input,
hidden and output layer.
2. The activation function that converts a neuron's
weighted input to its output activation.
3. The learning process for updating the weights of the
interconnections.
The main characteristic of ANN is self-learning without
prior knowledge of the complex non-linear relationships that
exist between the input and output variables. Another
advantage is that this type of approach also makes it possible
to use several predictor variables simultaneously.
2.2 Decision Tree
Decision tree [24] is one of the most popular and efficient
technique in data mining which is established and well-
explored by many researchers. Decision trees are categorized
as a supervised method that trying to find the relationship
between input attributes and target attributes which represent
the relationship in structure as a model. The model
constructed by using input attributes to predict target
However, some decision tree algorithms may produce a large
structure of tree size such as J48 which is an implementation
of C4.5 algorithm [25]. C4.5 was a version earlier algorithm
developed by J. Ross Quinlan.
3. PROPOSED METHODOLOGY
The overall work has two parts:
1. Data collection, preprocessing and preparation of database.
2. Application of AI based model to the dataset for validation
of the used techniques to predict the COVID-19 status i.e.
either the person is affected or not.
The publicly released COVID data form Israeli Ministry of
Health has been utilized for the preparation of the required
dataset. Next, this dataset has been used for both training and
validation for the Artificial Neural Network and Decision tree
on this classification problem. All these AI based models are
implemented using MATLAB 2018 in a laptop containing i3
processor and 4 GB processor. The details of the proposed
methodology have been elaborated in following sub-sections.
3.1 Data Collection and Preparation
The Israeli Ministry of Health [26], [27], [28] publicly
released data of individuals who were tested for SARS-CoV-2
via RT-PCR assay of a nasopharyngeal swab. The initial
dataset contains daily records of all the residents who were
tested for COVID-19 nationwide. Various information that
was provided in the datasheet includes clinically tested
symptoms like cough, fever, sore throat, shortness of breath,
headache. Based on these data, Artificial Neural Network and
Decision Tree model will be trained that will help to predict
COVID19 test results using eight binary features: gender, age
60 years or above, known contact with an infected individual,
and five initial clinical symptoms. The original data collected
for the months of March and April-2020. Then it was
processed to make it feasible for further analysis. A few steps
were executed to convert the large data into a clean data set.
The steps that were followed are as follows:
1. Data for a time period was segregated
2. Rows containing missing data were eliminated
3. Erroneous and wrong data were removed
4. Non binary data was converted to 0 and 1 in the
following manner
i. Gender: male -1; female -0
ii. Age above 60: yes-1; no-0
iii. Other information: contact with
confirmed-1; others-0
iv. Corona result:positive-1; negative-0
The modified training validation data set consisted of records
from 7,968 tested individuals (of whom 3214 were confirmed
to have COVID-19), from the period March 22th, 2020
through March 31st, 2020. The test set contained data from
the subsequent week, April 1st through April 7th (5,732 tested
individuals, of whom 1952 were confirmed to have COVID-
19).
The following list describes each of the dataset’s features used
by the model for training and testing:
A. Basic information:
1. Sex (male/female).
2. Age ≥60 years (true/false)
International Journal of Computer Applications (0975 – 8887)
Volume 184 – No.9, April 2022
9
B. Symptoms:
3. Cough (true/false).
4. Fever (true/false).
5. Sore throat (true/false).
6. Shortness of breath (true/false).
7. Headache (true/false).
C. Other information:
8. Known contact with an individual confirmed to have
COVID-19 (true/false).
3.2 Artificial Neural Network Based Model
To classify the COVID-19 data using ANN based model,
Multilayer Feed Forward Artificial Neural Network with one
hidden layer which consist of 20 nodes has been used. On the
other hand, the input layer consists of 8 nodes which are
different symptoms and features of COVID-19 patients.
Output layer of ANN model consists of only 1 node that
indicates the predicted value (COVID 1 or 0) i.e. the status of
the patients.
Initially, the training dataset is used to train the ANN model
with the use Back Propagation Algorithm. The trained ANN
model is then cross validated using the training data itself and
the performance and accuracy was noted. Cross-validation is a
resampling procedure used to evaluate machine learning
models on a limited data sample. Next, the trained model is
tested against a new dataset that contained inputs for the
period 1st April through 7th April, 2020. The performance
and accuracy are also noted for testing new case. All the
results are given in results section.
3.3 Decision Tree Based Model
C4.5 algorithm produces decision tree classification for a
given dataset by recursive division of the data. The decision
tree is grown using Depth-first strategy. On data testing, this
algorithm will emphasized on splitting dataset and by
selecting a test that will give best result in information gain.
With no of fold 10, and confidence factor 0.25, the algorithm
is implemented for construction of decision tree. Decision
Tree provides vary fast prediction but it has least accuracy
than other approaches. Decision Tree shows that one
directional path or inference for classification of new data
based on different attribute value. Results related with cross
validation and testing new data are given in next section.
4. RESULTS AND DISCUSSION
In this section, the detailed results corresponding to two cases
of AI models i.e. using ANN and Decision Tree to detect if a
person is affected be COVID or not are shown and discussed.
Following figure shows the Neural Network model (using
MATLAB) to detect if any person is COVID Positive or not,
based on some input symptoms and features.
Fig. 1: ANN training model for COVID detection
The Table 1 summarizes the classification accuracy for two
tested cases i.e. cross validation and testing new cases for
ANN.
Table 1: Accuracy for ANN model
Process
Percentage Accuracy
Cross Validation
84.80%
New Dataset
85.80%
The Table 2 shows the classification accuracy of decision tree
the two tested cases i.e. cross validation and testing new
cases.
Table 2: Accuracy for Decision Tree model
Process
Percentage Accuracy
Cross Validation
83.75%
New Dataset
85.79%
From above two tables, it has been observed that classification
accuracy is better for ANN over the decision tree for both
Cross validation and testing new cases. Hence, it can be stated
that the ANN is more suitable over decision tree in terms of
classification accuracy of COVID-19 dataset.
Figure 2 shows the output Decision Tree model for to detect if
any person is COVID-19 Positive or not.
International Journal of Computer Applications (0975 – 8887)
Volume 184 – No.9, April 2022
10
Fig. 2: Decision Tree model for COVID detection
Now, a comparison with respect to runtime of the ANN and
Decision Tree for this particular problem of classification of
COVID-19 has been observed. From Table 3, it can be clearly
noted that Decision Tree is very fast classification algorithm
compare to the ANN as it required only 38.18 sec for decision
tress training whereas ANN required almost 10 times.
Table 3: Runtime for ANN and Decision Tree
Process
Runtime (Sec)
ANN
363.61
Decision Tree
38.18
However, percentage of error for decision tree is slightly
higher than the ANN for both cross validation and testing new
case. Following Fig. 3 shows the percentage of error for
different cases during classification.
Fig. 3: Percentage of Error for different cases
Next, we shall show confusion matrix for the different case of
testing using ANN and Decision Tree. Table 4 and 5 shows
the confusion matrix for cross validation and testing new case
using ANN respectively.
Table 4: Confusion Matrix for Cross Validation Using
ANN
Target Value
Predicted
Value
0
1
0
4321
777
1
433
2437
Table 5: Confusion Matrix for Testing New Case Using
ANN
Target Value
Predicted
Value
0
1
0
3332
368
1
448
1584
Table 6 and 7 shows the confusion matrix for cross validation
and testing new case using decision tree respectively.
Table 6: Confusion Matrix for Cross Validation Using
Decision Tree
Target Value
Predicted
Value
0
1
0
4336
418
1
877
2337
15.2
16.25
14.2
14.21
ANN Cross
Validation
Decion Tree
Cross
validation
ANN Test New
Case
Decion Tree
Test New Case
Percentage of Error
International Journal of Computer Applications (0975 – 8887)
Volume 184 – No.9, April 2022
11
Table 7: Confusion Matrix for Testing New Case Using
Decision Tree
Target Value
Predicted
Value
0
1
0
3331
449
1
365
1587
From the above confusion matrix, it is very clear to us that
both ANN and Decision tree both not able to detect COVID
cases for few cases (i.e. False Negative or FN) and also they
detect few case as COVID positive by mistake (i.e. False
Positive or FP). As an example, for cross validation using
ANN, 777 numbers of FN and 433 numbers of FP are
detected which are not desirable. For this reason,
classification accuracy has been reduced to 84.80%. The
reason behind this kind of outputs is the asymptotic case for
COVID-19. As it is already known to us that many person do
not show any symptoms but they are affected by the COVID.
Hence, the proposed models also detect them as normal case
i.e. FNs are detected which of obvious a drawback of the AI
based automated diagnosis process. Similarly, it is also
possible that showing symptoms like fever; cough etc. does
not guarantee the COVID. The person may suffer due to
others disease. In this scenario, computer can predict them as
COVID positive case i.e. FPs are detected. However, the AI
based COVID-19 diagnosis is still very promising techniques
where medical facility like RTPCR test has limited
availability.
5. CONCLUSION
Here, the aim is to develop an automated diagnosis system for
COVID-19 using two popular AI based models namely
Artificial Neural Network and Decision Tree by observing the
symptoms and features of the respective patients. This paper
focuses on how ANN and decision tree can be used to detect
if any person is COVID positive or not, based on eight input
parameters. The dataset has been collected from Israeli
Ministry of Health website where all data are available
publicly for further analysis. In this study, both ANN and
decision tree model have been trained and tested for
validation of the proposed methodology. It has observed that
performance of the ANN is superior over decision tree in
terms of classification accuracy for both cross validation and
testing new cases. However, the runtime for decision tree is
very small compare to ANN. It is an advantage of using
decision tree for classification problem. It is possible to
achieve satisfactory classification accuracy for both models
although other advanced AI model or algorithm may be
applied to improve the performance in future. If healthcare
facility (i.e. RTPCR test) is limited, this AI based diagnostic
framework can be used where automated decision can be
taken by the computer or machine instantly without waiting
for the report from the pathologist.
6. REFERENCES
[1] World Health Organization. Report of the WHO-China
Joint Mission on Corona virus Disease 2019 (COVID-
19) Geneva: World Health Organization; 2020
https://www.who.int/docs/default-
source/coronaviruse/who-china-joint-mission-on-covid-
19-finalreport.
[2] Ministry of Health & Family Welfare Government of
India.
https://www.mohfw.gov.in/pdf/PreventionandManageme
ntofCOVID19FLWEnglish.pdf [Accessed on 20th May
2020].
[3] Liu J, Liao X, Qian S et al. Community transmission of
severe acute respiratory syndrome coronavirus 2,
Shenzhen, China, 2020. Emerg Infect Dis 2020
doi.org/10.3201/eid2606.200239
[4] Novel Corona Virus Map,
https://infographics.channelnewsasia.com/covid-
19/map.html [Accessed on 24th May 2020]
[5] Chung M, Bernheim A, Mei X, et al. CT Imaging
Features of 2019 Novel Coronavirus (2019-
nCoV). Radiology. 2020;295(1):202-207.
doi:10.1148/radiol.2020200230.
[6] Pan, Feng, et al. "Time course of lung changes on chest
CT during recovery from 2019 novel coronavirus
(COVID-19) pneumonia." Radiology (2020): 200370.
[7] Huang C, Wang Y, Li X, et al. Clinical features of
patients infected with 2019 novel coronavirus in Wuhan,
China. Lancet 2020; 395: 497–506.
[8] Liu Y, Yan LM, Wan L et al. Viral dynamics in mild and
severe cases of CVOID-19. Lancet Infect Dis
doi.org/10.1016/S1473-3099(20)30232-2
[9] Li, K., Fang, Y., Li, W. et al. CT image visual
quantitative evaluation and clinical classification of
coronavirus disease (COVID-19). Eur Radiol (2020).
https://doi.org/10.1007/s00330-020-06817-6.
[10] Aboul Ella Hassanien Sr., Lamia Nabil Mahdy Jr., Kadry
Ali Ezzat Jr., Haytham H. Elmousalami Jr.,Hassan Aboul
Ella Jr. Automatic X-ray COVID-19 Lung Image
Classification System based on Multi-Level
Thresholding and Support Vector Machine,
doi: https://doi.org/10.1101/2020.03.30.20047787.
[11] Singh, D., Kumar, V., Vaishali et al. Classification of
COVID-19 patients from chest CT images using multi-
objective differential evolution–based convolutional
neural networks. Eur J Clin Microbiol Infect Dis (2020).
https://doi.org/10.1007/s10096-020-03901-z.
[12] Yan, R. et al. Chest CT Severity Score: An Imaging Tool
for Assessing Severe COVID-19. Radiology:
Cardiothoracic Imaging 2020, 2(2).
https://doi.org/10.1148/ryct.2020200047.
[13] Rajinikanth, V., Nilanjan Dey, Alex Noel Joseph Raj,
Aboul Ella Hassanien, K. C. Santosh, and N. Raja.
"Harmony-search and otsu based system for coronavirus
disease (COVID-19) detection using lung CT scan
images." arXiv preprint, arXiv:2004.03431 (2020).
[14] S. Mandal, “Identification of Severity of Infection for
COVID-19 Affected Lungs Images using Elephant
Swarm Water Search Algorithm” International Journal of
Modelling and Simulation, 2021, doi:
10.1080/02286203.2021.1934797.
[15] A.M.U.D. Khanday et al., Machine learning based
approaches for detecting COVID-19 using clinical text
data, Int. J. Inf. Technol. (2020),
https://doi.org/10.1007/s41870-020-00495-9.
[16] C. Iwendi, et al., COVID-19 patient health prediction
using boosted random forest algorithm, Front. Public
Health 8 (2020),
https://doi.org/10.3389/fpubh.2020.00357.
International Journal of Computer Applications (0975 – 8887)
Volume 184 – No.9, April 2022
12
[17] X. Chen, Z. Liu, Early prediction of mortality risk among
severe COVID-19 patients using machine learning,
preprint, Epidemiology (2020),
https://doi.org/10.1101/2020.04.13.20064329.
[18] Carole H. Sudre et al, Attributes and predictors of Long-
COVID: analysis of COVID cases and their symptoms
collected by the COVID Symptoms Study App, medRxiv
preprint (2020) doi:
https://doi.org/10.1101/2020.10.19.20214494.
[19] Yazeed Zoabi et al, Machine learning-based prediction of
COVID-19 diagnosis based on symptoms, npj Digital
Medicine (2021) 4:3 ; https://doi.org/10.1038/s41746-
020-00372-6.
[20] Pijush Dutta, Shobhandeb Paul, Asok Kumar,
Comparative analysis of various supervised machine
learning techniques for diagnosis of COVID-19,
Electronic Devices, Circuits, and Systems for Biomedical
Applications (2021). https://doi.org/10.1016/B978-0-
323-85172-5.00020-4.
[21] Mei, X. et al. Artificial intelligence–enabled rapid
diagnosis of patients with COVID-19. Nat. Med. 26,
1224–1228 (2020).
[22] S. Mandal, G. Saha, and R. K. Pal, “A Comparative
Study on Disease Classification Using Different Soft
Computing Techniques”, The SIJ Transactions on
Computer Science Engineering & its Applications
(CSEA), vol. 1(3), pp. 59-66, 2014
[23] S. Mandal, G. Saha, and R. K. Pal, “An Approach
towards Automated Disease Diagnosis & Drug Design
Using Hybrid Rough-Decision Tree from Microarray
Dataset”, Journal of Computer Science and System
Biology, vol. 6(6), pp. 337-343, 2013,
DOI:10.4172/jcsb.1000130.
[24] S. Mandal, and I. Banerjee, “Cancer Classification Using
Neural Network”, International Journal of Emerging
Engineering Research and Technology, vol. 3(7), pp.
172-178, 2015.
[25] S. Mandal, G. Saha, and R. K. Pal, “Neural Network
Training Using Firefly Algorithm”, Global Journal on
Advancement in Engineering and Science, vol. 1(1), pp.
07-11, 2015.
[26] COVID-19-Government Data.
https://data.gov.il/dataset/covid-19 (2020).
[27] The Novel CoronavirusIsrael Ministry of Health.
https://govextra.gov.il/ministry-ofhealth/corona/corona-
virus-en/ (2020).
[28] COVID-19-Government Data Information.
https://data.gov.il/dataset/covid-19/resource/3f5c975e-
7196-454b-8c5b-ef85881f78db/download/-readme.pdf
(2020).
[29] Covid-19 India data information. https:
www.covid19india.org (2020)
IJCATM : www.ijcaonline.org
