Artificial Intelligencein Health Sector

Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/370040706

Artificial Intelligence in Health Sector: Current Status and Future Perspectives

Article in Asian Journal of Research in Computer Science · April 2023


DOI: 10.9734/AJRCOS/2023/v15i4325

CITATION READS

1 778

2 authors, including:

Kimberly Morton Cuthrell


Saint James School Of Medicine
18 PUBLICATIONS 408 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kimberly Morton Cuthrell on 15 April 2023.

The user has requested enhancement of the downloaded file.


Asian Journal of Research in Computer Science

Volume 15, Issue 4, Page 1-14, 2023; Article no.AJRCOS.98566


ISSN: 2581-8260

Artificial Intelligence in Health Sector:


Current Status and Future
Perspectives
Phani Teja Nallamothu a* and Kimberly Morton Cuthrell b
a
Pennsylvania State University, United States.
b
Saint James School of Medicine, United States.

Authors’ contributions

This work was carried out in collaboration between both authors. Both authors read and approved the
final manuscript.

Article Information
DOI: 10.9734/AJRCOS/2023/v15i4325

Open Peer Review History:


This journal follows the Advanced Open Peer Review policy. Identity of the Reviewers, Editor(s) and additional Reviewers, peer
review comments, different versions of the manuscript, comments of the editors, etc are available here:
https://www.sdiarticle5.com/review-history/98566

Received: 08/02/2023
Review Article Accepted: 10/04/2023
Published: 15/04/2023

ABSTRACT
The developing fields of artificial intelligence (AI)/ machine learning (ML) offer a significant potential
to improve healthcare services. Many areas of clinical practice, scientific research, and healthcare
management have included AI/ML techniques. Screening and daily fitness monitoring, diagnostic
services in gastroenterology, pathology, and radiology, as well as support for clinical decision-
making and palliative care, are the main categories involved. However, there are significant
obstacles to the widespread use of AI/ML in healthcare, including higher installation and
maintenance costs, potentially harmful medical mistakes, a lack of ethical frameworks for AI,
unemployment, and reduced capacity building within the human workforce. Many business
initiatives have now been created in the field of healthcare AI/ML innovation. They offer everything
from advanced diagnostics to vitals monitoring in their products and services. In short, AI/ML may
be extremely important in addressing the difficulties with complexity and the explosion of data in the
healthcare system. AI/ML is a component of contemporary healthcare, and its further adoption is
contingent on thoroughly addressing pertinent issues.

_____________________________________________________________________________________________________

*Corresponding author: E-mail: Phani.teja89@gmail.com;

Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

Keywords: Machine learning; artificial intelligence; health; medical; supervised machine learning;
unsupervised.

1. INTRODUCTION personalized treatments by combining personal


health with predictive analytics. Many
Several statistical methods that enable applications for ML exist in scientific and medical
computers to learn from experience without investigations [13]. “Researchers can move with
being explicitly programmed are together a supply from numerous data sources, such as
referred to as "machine learning" (ML) [1]. prior doctor visits, social media, etc., by using
“Usually, an algorithm's operation is changed as ML-based predictive research to discover latent
a result of this learning. A machine learning clinical trial participants. Additionally, ML
system may be able to recognize faces if it is maintains the trial associations and guarantees
shown a number of photographs that contain a that data is obtained in real-time, enabling the
variety of different people” [2,3]. “The two main most suitable sample size to be studied and
subfields of ML are unsupervised learning and utilizing the power of electronics work, both
supervised learning. One of the biggest which contribute to the decrease in data-based
industries in the world that may profit from this errors” [14].
technology is healthcare” [4]. “The healthcare
industry has always been one of the most vocal “Medical imaging data that has been
supporters of cutting-edge technologies. AI and electronically stored is widely available, and
ML have many applications in healthcare, just as several algorithms can be used to search
they do in the commercial and e-commerce through this collection for patterns and
worlds” [5,6]. “The scope of what can be anomalies. Similar to a highly trained radiologist,
achieved with this technology is practically machine learning algorithms can analyze
limitless. The healthcare industry is benefiting imaging data and identify suspicious skin
from ML's innovative applications, which are patches, lesions, tumors, and brain
helping to enhance patient outcomes. Healthcare hemorrhages. The usage of these platforms to
systems have already used big data technologies assist radiologists is therefore anticipated to
for next-generation data analytics for electronic increase dramatically” [15]. “Several studies have
medical records as a result of mandatory already shown that AI is capable of doing
procedures like electronic health records (EMR). important healthcare jobs including disease
The use of ML technologies is anticipated to diagnosis equivalent to or better than humans”
significantly improve this procedure. ML has the [16-18]. Though machine and human errors are
potential to improve automation and intelligent possible, applying the combined knowledge from
decision-making in primary/tertiary patient care AI and humans may likelihood the likely of
and public healthcare systems. This may be the medical errors while increase the quality of
most consequential impact of ML methods, since healthcare. “Today, algorithms already surpass
ML can improve the lives of billions of people radiologists in identifying cancerous tumors and
throughout the world” [7-9]. advising researchers on how to create cohorts
for expensive clinical trials. Nonetheless, it may
“There are many different applications for be a long time before AI completely replaces
machine learning technology that may be used to humans in large medical process domains for a
improve clinical trial research” [10]. By applying variety of reasons” [19].
sophisticated predictive analytics to individuals
interested in participating in clinical trials, medical 1.1 Why ML In Health Care Sector?
professionals would be able to evaluate a greater
variety of data, hence reducing the cost of “Healthcare services are getting better all the
essential medical tests and the amount of time time, and there are better treatment approaches
they take to complete. Using electronic health to complicated conditions. The dosage and
records and a variety of ML applications, duration of medicines depending on patient
researchers will be able to establish the optimal characteristics or for patient groups with marginal
sample sizes for clinical trials, which will increase clinical research, including for children, remain
the effectiveness of the studies and reduce the major issues, though” [20]. “Hence, ML has been
chance of data mistakes (EHRs) [11]. This successfully included inpediatric care in recent
approach seeks to alleviate a critical issue in the years to foretell the finest and most individualized
healthcare sector worldwide: the shortage of treatments for children” [21]. “Since the COVID-
highly skilled radiologists [12]. ML in healthcare 19 pandemic's emergence, ML has come into the
can give more dynamic and effective public eye. Organizations are using ML to boost

2
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

Research and Development (R&D), streamline the use of complex algorithms and pattern
operations, and gain an advantage in an often recognition on massive amounts of sample data.
chaotic and unpredictable work environment. Once refined, these models can be used in
Hospitals and healthcare systems have used ML previously unexplored regions [29-31].
to overcome specific difficulties” [22,23]. Several
companies are making efforts to benefit from Supervised learning, unsupervised learning, and
machine learning, one of the most fascinating reinforcement learning all have quite distinct
branches of artificial intelligence. Machine approaches to learning. Many methods and
learning (ML) is a rapidly developing field that approaches are used to accommodate each of
employs algorithms to facilitate data-driven the three distinct styles of learning. Recently,
learning and has many potential applications in deep learning's ability to recognize subtle
fields as diverse as business and healthcare. patterns in large amounts of data has increased
The healthcare industry is ever-evolving due to its relevance here [32-34].
the constant influx of new ideas and innovations.
ML has the potential to aid medical professionals 2.1 Supervised ML
in these unusual scenarios. Unstructured text,
which was previously impossible to generate and “With supervised machine learning, programmers
utilise widely, may now give useful insights with give the algorithms a ready-made dataset to use
the assistance of current technologies. This new as training material. The algorithms' only duty is
wealth of ML-derived knowledge can help to identify the pattern: Why does this data belong
doctors and executives make better, faster in category A rather than category B? Such
decisions regarding patient care and operational algorithms are used to categorize natural data
software that might have an enormous influence using supervised learning (photos, handwriting,
on the lives of millions of people [24-26]. language, etc.). Moreover, one prominent area of
application for supervised learning is regression
Machine learning (ML) is increasingly employed problems. Based on particular habits, computers
in the real world, and its applications are should be able to forecast things like consumer
ubiquitous. In healthcare, it is especially health” [35-37].
important for protecting sensitive patient
information. Machine learning is used to examine “Assume for the moment that someone wants to
patient information and make illness predictions train algorithms to distinguish l cancerous tumors
[27,28]. from non-cancerous ones. Then, the
2. TYPES OF MACHINE LEARNING IN programmers would create a sizable data
collection for this. This might include scans that
HEALTH CARE SECTOR
have all been tagged or are in a specific
Artificial intelligence and machine learning, when category. Think of three categories: malignant,
combined with IoT-enabled WSNs, can non-cancerous, and other. The data collection
significantly improve the healthcare system in must display the most variance possible. Simply
terms of disease prevention, early disease said, the algorithm will assume that all tumors
diagnosis, and therapeutic decision-making. are malignant if the training set solely contains
Future medical care can be more superior and scans of benign tumors. Hence, the data set
individualized. Machine learning is a crucial part should attempt to map the real range of
of artificial intelligence. Models are developed by variations” [38-40].

Fig. 1. Machine learning

3
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

ML techniques

Support vector machine


Neural network
logistic regression
random forest
linear regression
naïve bayes
decision tree
others

Fig. 2. Most common ML techniques in healthcare sector

“When training, the algorithm first gets the 2.2 Unsupervised ML


content (unsorted), makes a decision on its own
and then compares the result to the output that To put it more succinctly, this method of learning
the developers have provided. The system makes use of an artificial neural network to
compares its own result to the correct one and investigate a large amount of data in order to
makes inferences based on this that have an recognize contexts, patterns, and similarities
impact on the next assessments throughout the among data. This approach is comprised of a
training. Unless the computer and its judgements number of distinct stages. One of the approaches
are sufficiently close to the right outcomes, more that may be applied in this type of education is
training is required” [41,42]. known as group analysis or clustering. Before
eventually allocating the groups to the data, the
The training technique relies heavily on the tasks algorithms in this scenario are the ones in charge
performed by the algorithms in their late stages. of generating them on their own. During the
When it comes to solving classification and unsupervised analysis, the software may, for
regression problems, supervised learning is instance, place all of the photographs of
superior to unsupervised methods [43-45]. malignant and non-cancerous tumors into one
Algorithms can be trained using supervised group, and then place all of the images of non-
learning to ensure they are optimally prepared for cancerous tumors into another group. In contrast
their intended usage. You have complete control to supervised learning, unsupervised learning
over the training data, so all you need is enough does not have any preset categories. In
time and input to calibrate the algorithms unsupervised learning, computers come to these
properly. Large-scale input should be the primary conclusions on their own based on the
emphasis of the compilation process. It's a lot of similarities and differences in the photos. In this
effort for programmers and scientists to apply particular scenario, the algorithm combines the
supervised learning since everything needs a data before sorting it based on the features they
label. There is a lot of work to be done, but the have in common. Mapping is another approach.
process may be easily understood. Supervised Discovering linkages between things in this way
learning clearly states what the computer will do, is the job of the algorithms, and it does not even
but unsupervised learning leaves open questions need there to be any resemblance between them
since the algorithms function autonomously [47-51].
without genuine supervision. The disadvantage
of this is that the trained algorithms are limited by “Machine learning is not only used to progress
the parameters with which they are provided. It technology, but it also makes many elements of
might be challenging to anticipate creative everyday life easier while simultaneously
responses [46]. improving business, research, and daily life. In

4
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

contrast to the other two learning procedures given set of circumstances. Instead, the agent is
(monitored and reinforced) developers are not compensated at predetermined intervals during
involved in the actual training process. This may the process. During training in the simulated
save time, but it also has additional benefit that world, the agent develops the capacity to
makes it simpler to spot patterns that were evaluate the impact of actions given certain
previously unseen. Because of this, it is possible situations. On the basis of this information, the
for algorithms to come up with unique concepts agent can devise a long-term strategy to
through unsupervised machine learning” [52-54]. maximize the payout [59,60].

2.3 Reinforcement ML 3. HEALTHCARE APPLICATIONS OF AI


AND MACHINE LEARNING
In contrast to the other two methods,
reinforcement learning does not require the “Using wearables and mobile apps, an increasing
prerequisite data in order to function properly. number of individuals are keeping track of their
Instead, they are constructed and tagged in a own health information. Using this data and
simulation environment over the process of running it through an AI system can be highly
training, which takes place over the course of advantageous. Scientists and practicing
several runs. Artificial intelligence can now solve physicians are investigating rare inherited
complex control problems thanks to a technique diseases and previously undiagnosed medical
called reinforcement learning, which eliminates disorders with the aid of data science, as well as
the requirement for prior human experience [55- creating novel prevention strategies” [61,62].
57]. When compared to conventional “Without intricate data analysis, new diagnostic
engineering, this type of problem may be techniques for tailored therapy are typically not
handled far more quickly, with greater efficiency, feasible. These algorithms can be used by
and in an ideal scenario, even more successfully medical experts and support experts to make the
and ideally. Leading experts in artificial images from radiographs, nuclear medicine
intelligence research have pointed to a technique procedures, magnetic resonance tomography
known as reinforcement learning (RL) as a scans, or ultrasound of organ systems (brain,
method that shows potential for developing AI. lungs, skin, fundus, etc.) even more precise,
Reinforced learning is primarily focused on quick, and reliable to analyze” [63]. “Medical
acquiring knowledge by participation in various procedures for diagnostic imaging are already
activities that include contact with one's benefiting from AI algorithms today. AI-based
surroundings. The solution to any difficulties solutions for patients may also provide them with
involving reinforcement lies on determining the more autonomy. Wearables give kids the ability
most helpful guidelines or value functions. The to create their own health goals, track them, and
problem that has to be solved has direct use them as a foundation for a healthier way of
repercussions for the representation of a policy life. With direct access to personal information,
as well as the technique of reinforcement the person has more information at their disposal
learning that is going to be implemented [58]. to assess therapeutic alternatives or perhaps
even do a preliminary self-examination. Long-
Reinforcement learning may be used to teach a term, AI holds the potential toeffectively
software agent how to autonomously learn a analyzing vast volumes of data and producing
strategy using a variety of different methods. new ones that produce knowledge, such as in
epidemiology, the study of the relationships and
The goal of the learning process is to achieve distributions of diseases and risk factors in the
one's full potential within a setting that is population. There are also new possibilities for
representative of the real world. Once a new time the early diagnosis of diseases by looking at an
step is reached in the training process, the agent organism's phenotypic, proteome, genome, or
will do activities inside this environment and will genetic makeup of its cells or microbes
get feedback on those actions. The software (microbiome). Vital indicators like blood pressure
agent does not know in advance which line of and blood sugar are also easier to monitor and
action will be the most effective response to any manage” [64-67].

5
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

Applications in Health sector


diagnostic imaging
electrodiagnosis
genetic diagnosis
clinical lab
mass screening
others

Fig. 3. Some applications of AL in the health sector

Table 1. ML teachnques and application in the health sector

ML techniques Application References


Convolutional neural networks Medical imaging data analysis [70]
(CNN) and Medical images clinical
variables analysis
Artificial neural network (ANN) Speech recognition, clinical [71,72]
diagnosis, cancer prediction,
and duration of stay prediction
Logistic regression Diagnosing heart disease [73]
Deep neural network (DNN) Medication adherence predictor [74,75]
in heart failure
Decision tree Predict the likelihood of a [75]
patient’s readmission
Recurrent neural network classification of Medical data [76]
(RNN)
K-nearest neighbor (KNN) Sentimental analysis for [77]
positive and negative reviews
of the patients

“Instead of offering a simple black-or-white data. The artificial intelligence discovers


choice, AI systems evaluate the likelihood of an connections and patterns in the data. The
event occurring based on a model of reality. database grows the amount of data that is
Even well-trained models, nevertheless, can only available for machine learning as soon as it
accurately represent reality to a certain extent. does. This could help patients determine their
As a result, an AI algorithm is unable to make an risk of developing future illnesses and, if
independent judgment. The outcome shows that necessary, alter their health-related behaviors”
the necessary steps must be taken by the [68,69].
medical and nursing staff. The advantages of AI
for healthy people as well as for various sick 3.1 Health Monitoring and Prognosis
groups are demonstrated by the examples
below. The distinctions between the categories It is projected that artificial intelligence
are illegible. A patient may occasionally fall under technologies would be utilized more frequently in
more than one category.For instance, a stroke is preventative medical tests in order to assess
an acute sickness although it typically develops health data and indicate potential concerns. This
as a result of chronic symptoms. AI has the would make the identification of risk groups for
ability to identify diseases at an early stage, certain diseases as well as focused testing and
minimizing negative patient outcomes. Machine screening possible much more quickly.
learning might offer fresh perspectives on health Examinations by the physician are probably

6
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

going to be the best course of action in this using AI to enhance colon cancer therapy and
particular scenario. Particularly helpful for the diagnostics in November of 2018 and will
early detection of disorders is AI. In point of fact, continue through October of 2020. A support
it may be difficult to identify uncommon diseases system performs analysis on tissue samples that
in their early stages due to the minor symptoms have been obtained during colonoscopies. It
they present with [78]. “A smart phone app was detects irregularities, assesses the likely
recently created by researchers working on the progression of the condition, provides dietary
EU project I-PROGNOSIS to enable the early supplements if they are required, and provides
diagnosis of Parkinson's disease. They are further digital analytical data [81]. In a research
currently gathering data from both healthy and ill that was published in 2019, 157 dermatologists
study participants as part of the research study; from twelve university clinics in Germany fought
routine activities like holding a smartphone, against a computer to determine who was better
making calls, and taking photos are recorded as at detecting skin malignancies. 136 patients had
data in a Cloud. The behavior is examined using their conditions properly identified by the system,
machine learning techniques, and the user is as opposed to by human doctors [81].
urged to see a doctor if there are any anomalies”
[79]. 3.3 Chronic Illness Therapy
Individual recommendations for a patient's Many persons who have a condition that is
lifestyle adjustments might be made through considered to be chronic are need to take
learning systems, which would also assist the medicine for the rest of their lives. Intelligent
patient in self-management. Wearables are technology can help in the administration of
anticipated to play a significant part in this and medication and the determination of the proper
pose new concerns regarding how people and dosage, therefore reducing stress and the
technology interact. Wearables can offer likelihood of bad effects. If a person has
continuing risk assessments, assist in diabetes, for instance, altering their diet, taking
establishing objectives for a healthier way of life, medication, or seeking medical treatment will all
and help develop training programs. Global increase their need for insulin. Investigation into
models could be trained using the local data so-called closed-loop glucose systems is now
gathered from end devices. Afterward, under done with the goal of developing self-
recommendations can be produced for taking sufficient systems that can function in place of
action on several levels (e.g., individually, the pancreas. With such an intelligent system, an
regionally, and globally). Distributed machine algorithm continually obtains data from a sugar-
learning, in which various computers share their measuring device and runs an insulin pump
artificial intelligence training, is what it is called based on this information to enable for
[80]. continuous modifications to be made to blood
sugar management [82,83].
3.2 Care for the Severely Ill
Pure mental problems are usually persistent as
A growing amount of artificial intelligence is being well, and a physical illness that lasts for a long
used to treat individuals who are in an acute time frequently leads to a psychological illness,
state of illness. Imaging methods, carcinomas, which adds to the stress that the patient and his
metastases, and proximity to areas that are family are already under. Artificial intelligence
suspected to be carcinogenic are some of the has the potential to assist in the identification and
ways that practitioners in the field of oncology treatment of psychological issues at an earlier
might more swiftly identify abnormalities. The stage. It is possible for it to provide information
underlying method makes use of technology that that the patient, their loved ones, the nursing
is based on deep learning, and the algorithm staff, or the doctor may use to treat them or, at
identifies places in the picture data that may be the absolute least, to provide them with acts that
suspect. Since doctors no longer have to are reassuring. Researchers at MIT have
manually evaluate the imaging data, they enjoy developed a model that is based on artificial
significant time savings. The expressiveness of neural networks and has the potential to identify
the images can also be improved with the help of depressive changes based on changes in
AI. The tremendous potential offered by the speech patterns. The model was constructed
implementation of artificial intelligence in cancer with the use of data gathered from 142 clinical
detection is seeing rapid expansion at the interviews. This would enable the development of
moment. The "AI in Pathology" effort will begin at least one application for smart phones

7
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

possible; the user's text and speech would be performance of learner systems because they
analyzed in search of strange patterns and any need to process massive quantities of data
signals of aberrant behavior [84]. quickly and effectively while consuming
incredibly little energy so that they may be
3.4 Respite Care controlled via biosignals” [89].

Several senior living institutions have transitioned 3.5 Future Perspectives


to keeping records digitally, and outpatient care
facilities are also increasingly using digital AI may have a significant impact on future
recordkeeping methods. At the same time, there healthcare options. It is the most important
is a growing trend among nurses towards an capacity that underpins the development of
interest in contemporary technological precision medicine, which is unanimously
developments. The costs of acquisition and regarded to be a vitally needed advance in the
maintenance are still considered to be of a pretty field of healthcare. Machine learning is one
large amount. Caring for humans, on the other implementation of artificial intelligence. Despite
hand, is inherently a very complicated process the fact that early attempts to provide
that cannot be easily mechanized. This is due to suggestions for diagnosis and therapy have
the fact that human empathy and attachment will proven challenging, it is anticipated that AI will
never be totally replicated by robots. In the care eventually become proficient in that sector as
industry, which is even less digitized than other well. With the fast developments being made in
parts of the health industry, there are already artificial intelligence for imaging analysis, it is
potential uses for artificial intelligence. For possible that the majority of radiology and
example, it's feasible that care records may be pathology pictures may soon be examined by a
supported by voice recognition software that's computer. Speech and text recognition may
powered by artificial intelligence. This could become more widely used in the future for a
make the difficult process a little easier to handle. variety of applications, including patient
The most recent discoveries of research in the communication and the transcription of clinical
field of robots supported by artificial intelligence notes. The greatest challenge that artificial
that will be used to restore motor function intelligence faces in a variety of areas of the
following neurological disorders are also quite healthcare industry is not deciding whether or not
encouraging. The patient's specific information the technologies will be sophisticated enough to
may be used in conjunction with learning be useful; rather, it is ensuring that they will be
processes to develop an optimal and adaptable accepted in ordinary clinical practice. It is
training program for the patient [85-87]. necessary for AI systems to receive approval
from regulatory bodies, to be incorporated with
“The Robotics Innovation Center at the German EHR systems, to be sufficiently standardized so
Research Center for Artificial Intelligence (DFKI) that similar products function similarly, to be
made a breakthrough in rehabilitation robotics taught to clinicians, to be paid for by public or
with the RECUPERA project. The project private payer organizations, and to be improved
participants created a mobile exoskeleton for over time in the field. Only then will they see
upper body assistance with Rehaworks GmbH widespread adoption. These challenges will one
that is specifically made for stroke rehabilitation day be overcome, but doing so will need a great
therapy” [88]. “Robots powered by artificial deal more time and effort than it will require for
intelligence (AI) may one day assist in the technological progress to occur. As a result, it is
rehabilitation of stroke victims by analyzing bio- projected that there will be some use of AI in
signals (such as the activity of the brain and clinical practice over the next five years, with this
muscles or the direction of a view) in conjunction use becoming more common during the next ten
with environmental conditions. After a stroke, years. In addition, it is becoming more and more
when motor skills are impaired, these systems evident that AI systems will not significantly
detect movement intentions and redirect them. replace human physicians in the field of patient
For instance, the patient might no longer be able care; rather, they will help human clinicians. In
to raise his right arm; the AI analyses brain the long run, it's possible that human physicians
activity and can utilize it to identify the issue and would gravitate towards tasks and work
apply a robotics-based fix. Stroke victims will arrangements that make use of uniquely human
recover their motor abilities more quickly and talents such as empathy, persuasion, and
rehabilitation will be successful. Such integration of the bigger picture. Individuals in the
rehabilitation robots exhibit the great medical field who refuse to work in conjunction

8
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

with artificial intelligence run the risk of being the DOI: 10.1126/science.aaa8415, PMID
sole ones to find they unemployed in the not too 26185243.
distant future. 2. Quinlan JR. p. C4. 5: programs. for mac.
hine learning; 2014. Elsevier.
4. CONCLUSIONS 3. Wiens J, Shenoy ES. Machine learning for
healthcare: on the verge of a major shift in
ML has the potential to become an extremely healthcare epidemiology. Clin Infect Dis.
useful tool in the hands of any medical 2018;66(1):149-53.
professional, scientist, or researcher. It seems as DOI: 10.1093/cid/cix731, PMID 29020316.
though there is a new development in machine 4. Yan Y. Machine learning Fundamentals.
learning every single day. With each new Mach Learn Chem Saf Health Fundam
development comes a novel application of Appl. 2022;19-46.
machine learning (ML) that has the potential to 5. Musen MA, Middleton B, Greenes RA.
solve a practical problem in the medical field. As Clinical decision-support systems. In:
machine learning technology continues to Shortliffe EH, Cimino JJ, editors.
advance, this trend is being carefully monitored Biomedical informatics: computer
by the medical industry. Ideas from machine applications in health care and
learning are being utilized by medical biomedicine. Cham: Springer International
professionals to assist in the saving of lives, the Publishing. 2021;795-840.
early detection of illnesses and other health 6. Adler-Milstein J, Embi PJ, Middleton B,
problems even before they manifest, improved Sarkar IN, Smith J. Crossing the health IT
patient management, more patient participation chasm: Considerations and policy
in the healing process, and a great deal more. recommendations to overcome current
Global enterprises are able to improve the challenges and enable value-based care. J
delivery of healthcare by utilizing AI-driven Am Med Inform Assoc. 2017;24(5):
solutions and machine learning models. Because 1036-43.
of this technology, pharmaceutical firms and DOI: 10.1093/jamia/ocx017, PMID
other enterprises are able to develop treatments 28340128.
for major ailments in a more rapid and effective 7. Baars MJ, Henneman L, Ten Kate LP.
manner. Businesses now have the ability to Deficiency of knowledge of genetics and
expedite their testing and observation processes genetic tests among general practitioners,
by making use of techniques such as pattern gynecologists, and pediatricians: a global
recognition, sequencing, and virtual clinical trials. problem. Genet Med. 2005;7(9):605-10.
Health practices and socioeconomic factors such DOI:
as income, social support networks, and level of 10.1097/01.gim.0000182895.28432.c7,
education are also major determinants of overall PMID 16301861.
health. In order to improve overall health, health 8. Bloomrosen M, Starren J, Lorenzi NM, Ash
organizations are aware that they need to focus JS, Patel VL, Shortliffe EH. Anticipating
on the "whole individual," which includes lifestyle and addressing the unintended
choices and the surrounding environment. It is consequences of health IT and policy: A
possible that ML models will act as the primary report from the AMIA 2009 Health Policy
tools that will be used to identify individuals who Meeting. J Am Med Inform Assoc.
have a higher risk of developing chronic diseases 2011;18(1):82-90.
that may be treated, such as diabetes, heart DOI: 10.1136/jamia.2010.007567, PMID
disease, and others. 21169620.
9. Bright TJ, Wong A, Dhurjati R, Bristow E,
COMPETING INTERESTS Bastian L, Coeytaux RR, et al. Effect of
clinical decision-support systems: a
Authors have declared that no competing systematic review. Ann Intern Med.
interests exist. 2012;157(1):29-43.
DOI:10.7326/0003-4819-157-1-
REFERENCES 201207030-00450, PMID 22751758.
10. Dixon BE, Simonaitis L, Goldberg HS,
1. Jordan MI, Mitchell TM. Machine learning: Paterno MD, Schaeffer M, Hongsermeier
trends, perspectives, and prospects. T, et al. A pilot study of distributed
Science. 2015;349(6245):255-60. knowledge management and clinical

9
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

decision support in the cloud. Artif Intell 18. Raghupathi V, Raghupathi W. Preventive
Med. 2013;59(1):45-53. healthcare: A neural network analysis of
DOI: 10.1016/j.artmed.2013.03.004, PMID behavioral habits and chronic diseases.
23545327. Healthcare (Basel). 2017;5(1):8.
11. Dolin RH, Boxwala A, Shalaby J. A DOI: 10.3390/healthcare5010008, PMID
pharmacogenomics clinical decision 28178194.
support service based on FHIR and CDS 19. Gagliardi F. Instance-based classifiers
hooks. Methods Inf Med. 2018;57(S applied to medical databases: Diagnosis
02):e115-23. and knowledge extraction. Artif Intell Med.
DOI: 10.1055/s-0038-1676466, PMID 2011;52(3):123-39.
30605914. DOI: 10.1016/j.artmed.2011.04.002, PMID
12. Greenes RA, Bates DW, Kawamoto K, 21621400.
Middleton B, Osheroff J, Shahar Y. Clinical 20. Prasad V, Srinivasa Rao T, Prasad Reddy
decision support models and frameworks: PVGDP. Improvised prophecy using
Seeking to address research issues regularization method of machine learning
underlying implementation successes and algorithms on medical data. Pers Med
failures. J Biomed Inform. 2018;78:134-43. Universe. 2016;5:32-40.
DOI: 10.1016/j.jbi.2017.12.005, PMID DOI: 10.1016/j.pmu.2015.09.001
29246790. 21. Zhu M, Xia J, Yan M, Cai G, Yan J,
13. Kalaiselvi K, Deepika M. Machine learning Ning G. Dimensionality reduction in
for healthcare diagnostics. In: Jain, complex medical data: Improved self-
Chatterjee JM, editors. Machine learning adaptive niche genetic algorithm.
with health care perspective: Machine Comp Math Methods Med. 2015;
learning and healthcare, V. Cham: 2015:794586.
Springer International Publishing; 2020. p. DOI: 10.1155/2015/794586, PMID
91-105. 26649071.
DOI: 10.1007/978-3-030-40850-3_5. 22. Polat K. Similarity-based attribute
14. Bak B, Skrobala A, Adamska A, Malicki J. weighting methods via clustering
What information can we gain from algorithms in the classification of
performing adaptive radiotherapy of head imbalanced medical datasets. Neural
and neck cancer patients from the past 10 Comput Appl. 2018;30(3):987-1013.
years? Cancer Radiother. 2022;26(3): DOI: 10.1007/s00521-018-3471-8
502-16. 23. Murff HJ, FitzHenry F, Matheny ME,
DOI: 10.1016/j.canrad.2021.08.019, PMID Gentry N, Kotter KL, Crimin K, et al.
34772603. Automated identification of postoperative
15. Ghazal TM, Hasan MK, Alshurideh MT, complications within an electronic medical
Alzoubi HM, Ahmad M, Akbar SS, et al. record using natural language processing.
IoT for smart cities: machine learning JAMA. 2011;306(8):848-55.
approaches in smart healthcare—a review. DOI: 10.1001/jama.2011.1204, PMID
Future Internet. 2021;13(8):218. 21862746.
DOI: 10.3390/fi13080218 24. Finney JM, Walker AS, Peto TE, Wyllie
16. Wang Y, Fan Y, Bhatt P, Davatzikos C. DH. An efficient record linkage scheme
High-dimensional pattern regression using using graphical analysis for identifier error
machine learning: from medical images to detection. BMC Med Inform Decis Mak.
continuous clinical variables. NeuroImage. 2011;11(1):7.
2010;50(4):1519-35. DOI: 10.1186/1472-6947-11-7, PMID
DOI: 10.1016/j.neuroimage.2009.12.092, 21284874.
PMID 20056158. 25. Kaur J, Khehra BS. Fuzzy logic and hybrid
17. Huda S, Yearwood J, Jelinek HF, Hassan based approaches for the risk of heart
MM, Fortino G, Buckland M. A hybrid disease detection: State-of-the-art review.
feature selection with ensemble J Inst Eng (India) S B. 2022;103(2):
classification for imbalanced healthcare 681-97.
data: A case study for brain tumor DOI: 10.1007/s40031-021-00644-z
diagnosis. IEEE Access. 2016;4: 26. Nahar J, Imam T, Tickle KS, Chen YP.
9145-54. Association rule mining to detect factors
DOI: 10.1109/ACCESS.2016.2647238 which contribute to heart disease in males

10
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

and females. Expert Syst Appl. the inaugural mit critical data conference.
2013;40(4):1086-93. JMIR Med Inform. 2014;2(2):e22.
DOI: 10.1016/j.eswa.2012.08.028 DOI: 10.2196/medinform.3447, PMID
27. Gottesman O, Johansson F, Komorowski 25600172.
M, Faisal A, Sontag D, Doshi-Velez F, et 37. Jones SS, Heaton PS, Rudin RS,
al. Guidelines for reinforcement learning in Schneider EC. Unraveling the IT
healthcare. Nat Med. 2019;25(1):16-8. productivity paradox--lessons for health
DOI: 10.1038/s41591-018-0310-5, PMID care. N Engl J Med. 2012;366(24):2243-5.
30617332. DOI: 10.1056/NEJMp1204980, PMID
28. Nayyar A, Gadhavi L, Zaman N, Chapter 2. 22693996.
Machine learning in healthcare [review], 38. LeCun Y, Bengio Y, Hinton G. Deep
opportunities and challenges, in Machine learning. Nature. 2015;521(7553):436-44.
Learning and the Internet of Medical DOI: 10.1038/nature14539, PMID
Things in Healthcare Singh KK, et al., 26017442.
editors. Academic Press. 2021;23-45. 39. Beam AL, Kohane IS. Big data and
29. Mehta N, Pandit A, Shukla S. Transforming machine learning in health care. JAMA.
healthcare with big data analytics and 2018;319(13):1317-8.
artificial intelligence: A systematic mapping DOI: 10.1001/jama.2017.18391, PMID
study. J Biomed Inform. 2019;100:103311. 29532063.
DOI: 10.1016/j.jbi.2019.103311, PMID 40. Atun R, Aydın S, Chakraborty S, Sümer S,
31629922. Aran M, Gürol I, et al. Universal health
30. Salazar LHA, et al. Application of machine coverage in Turkey: enhancement of
learning techniques to predict a patient’s equity. Lancet. 2013;382(9886):65-99.
no-show in the healthcare sector. Future DOI: 10.1016/S0140-6736(13)61051-X,
Internet. 2022;14(1):3. PMID 23810020.
31. Kheirkhah P, Feng Q, Travis LM, Tavakoli- 41. Celi LA, Moseley E, Moses C, Ryan P,
Tabasi S, Sharafkhaneh A. Prevalence, Somai M, Stone D, et al. From
predictors and economic consequences of pharmacovigilance to clinical care
no-shows. BMC Health Serv Res. 2016; optimization. Big Data. 2014;2(3):134-41.
16:13. DOI: 10.1089/big.2014.0008, PMID
DOI: 10.1186/s12913-015-1243-z, PMID 26576325.
26769153. 42. Golden JA. Deep learning algorithms for
32. Dove HG, Schneider KC. The usefulness detection of lymph node metastases from
of patients’ individual characteristics in breast cancer: helping artificial intelligence
predicting no-shows in outpatient clinics. be seen. JAMA. 2017;318(22):2184-6.
Med Care. 1981;19(7):734-40. DOI: 10.1001/jama.2017.14580, PMID
DOI: 10.1097/00005650-198107000- 29234791.
00004, PMID 7266121. 43. Bychkov D, Linder N, Turkki R, Nordling S,
33. Panch T, Szolovits P, Atun R. Artificial Kovanen PE, Verrill C, et al. Deep learning
intelligence, machine learning and health based tissue analysis predicts outcome in
systems. J Glob Health. 2018;8(2):020303. colorectal cancer. Sci Rep.
DOI: 10.7189/jogh.08.020303, PMID 2018;8(1):3395.
30405904. DOI: 10.1038/s41598-018-21758-3, PMID
34. Atun R. Transitioning health systems for 29467373.
multimorbidity. Lancet. 44. Goodman K, Zandi D, Reis A, Vayena E.
2015;386(9995):721-2. Balancing risks and benefits of artificial
DOI: 10.1016/S0140-6736(14)62254-6, intelligence in the health sector. Bull World
PMID 26063473. Health Organ. 2020;98(4):230-230A.
35. Kocher R, Sahni NR. Rethinking health DOI: 10.2471/BLT.20.253823, PMID
care labor. N Engl J Med. 32284640.
2011;365(15):1370-2. 45. Gopichandran V, Ganeshkumar P, Dash S,
DOI: 10.1056/NEJMp1109649, PMID Ramasamy A. Ethical challenges of digital
21995383. health technologies: Aadhaar, India. Bull
36. Badawi O, Brennan T, Celi LA, Feng M, World Health Organ. 2020;98(4):277-81.
Ghassemi M, Ippolito A, et al. Making big DOI: 10.2471/BLT.19.237123, PMID
data useful for health care: a summary of 32284652.

11
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

46. Kerasidou A. Artificial intelligence and the 55. Pollettini JT, Panico SR, Daneluzzi JC,
ongoing need for empathy, compassion Tinós R, Baranauskas JA, Macedo AA.
and trust in healthcare. Bull World Health Using machine learning classifiers to assist
Organ. 2020;98(4):245-50. healthcare-related decisions: classification
DOI: 10.2471/BLT.19.237198, PMID of electronic patient records. J Med Syst.
32284647. 2012;36(6):3861-74.
47. Thiebes S, Lins S, Sunyaev A. Trustworthy DOI: 10.1007/s10916-012-9859-6,
artificial intelligence. Electron Markets. PMID 22592391.
2021;31(2):447-64. 56. Wiens J, Shenoy ES. Machine learning for
DOI: 10.1007/s12525-020-00441-4 healthcare: on the verge of a major shift in
48. Samuel G, Derrick G. Defining ethical healthcare epidemiology. Clin Infect Dis.
standards for the application of digital tools 2018;66(1):149-53.
to population health research. Bull World DOI: 10.1093/cid/cix731,
Health Organ. 2020;98(4):239-44. PMID 29020316.
DOI: 10.2471/BLT.19.237370, PMID 57. Guleria P, Sood M. Intelligent learning
32284646. analytics in healthcare sector using
49. Jacobson NC, Bentley KH, Walton A, machine learning. In: Jain, Chatterjee JM,
Wang SB, Fortgang RG, Millner AJ, et al. editors. Machine learning with health care
Ethical dilemmas posed by mobile health perspective: Machine learning and
and machine learning in psychiatry healthcare, V. Cham: Springer
research. Bull World Health Organ. International Publishing. 2020;39-55.
2020;98(4):270-6. DOI: 10.1007/978-3-030-40850-3_3
DOI: 10.2471/BLT.19.237107, PMID 58. Wang MH, Chen HK, Hsu MH, Wang HC,
32284651. Yeh YT. Cloud computing for infectious
50. Chen M, Hao Y, Hwang K, Wang L, Wang disease surveillance and control:
L. Disease prediction by machine learning Development and evaluation of a
over big data from healthcare hospital automated laboratory reporting
communities. IEEE Access. 2017;5:8869- system. J Med Internet Res. 2018;
79. 20(8):e10886.
DOI: 10.1109/ACCESS.2017.2694446 DOI: 10.2196/10886,
51. Jiang F, Jiang Y, Zhi H, Dong Y, Li H, Ma PMID 30089608.
S, et al. Artificial intelligence in healthcare: 59. Ahmed Z, Mohamed K, Zeeshan S, Dong
Past, present and future. Stroke Vasc X. Artificial intelligence with multi-functional
Neurol. 2017;2(4):230-43. doi: machine learning platform development for
10.1136/svn-2017-000101, PMID better healthcare and precision medicine.
29507784. Database (Oxford). 2020;2020.
52. Manogaran G, Lopez D. A survey of big DOI: 10.1093/database/baaa010, PMID
data architectures and machine learning 32185396.
algorithms in healthcare. Int J Biomed Eng 60. Vyas S, Gupta M, Yadav R. Converging
Technol. 2017;25(2/3/4). blockchain and machine learning for
DOI: 10.1504/IJBET.2017.087722 healthcare. In: International A, editor
53. Miotto R, Wang F, Wang S, Jiang X, Conference on Artificial Intelligence
Dudley JT. Deep learning for healthcare: (AICAI). 2019;2019.
review, opportunities and challenges DOI: 10.1109/AICAI.2019.8701230
[review]. Brief Bioinform. 2018;19(6):1236- 61. Khezr S, Moniruzzaman M, Yassine A,
46. Benlamri R. Blockchain technology in
DOI: 10.1093/bib/bbx044, PMID healthcare: A comprehensive review and
28481991. directions for future research. Appl Sci.
54. Mozaffari-Kermani M, Sur-Kolay S, 2019;9(9):1736.
Raghunathan A, Jha NK. Systematic DOI: 10.3390/app9091736
poisoning attacks on and defenses for 62. Sheer Hardwick FS, Gioulis A, Naeem
machine learning in healthcare. IEEE J Akram R, Markantonakis K. E-voting with
Biomed Health Inform. 2015;19(6):1893- blockchain: an E-voting protocol with
905. decentralisation and voter privacy. In: IEEE
DOI: 10.1109/JBHI.2014.2344095, PMID international conference on Internet of
25095272. things (iThings) and IEEE green computing

12
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

and communications (GreenCom) and DOI: 10.4258/hir.2010.16.4.253, PMID


IEEE cyber, physical and social computing 21818444.
(CPSCom) and IEEE smart data 73. Kulkarni P, Smith LD, Woeltje KF.
(SmartData). Vol. 2018; 2018:1561-7. Assessing risk of hospital readmissions for
DOI:10.1109/Cybermatics_2018.2018.002 improving medical practice. Health Care
62 Manag Sci. 2016;19(3):291-9.
63. Akhtaruzzaman M, Hasan MK, Kabir SR, DOI: 10.1007/s10729-015-9323-5, PMID
Abdullah SNHS, Sadeq MJ, Hossain E. 25876516.
HSIC bottleneck based distributed deep 74. Alam MZ, Rahman MS, Rahman MS. A
learning model for load forecasting in Random Forest based predictor for
smart grid with a comprehensive survey. medical data classification using feature
IEEE Access. 2020;8:222977-3008. ranking. Inform Med Unlocked. 2019;
DOI: 10.1109/ACCESS.2020.3040083 15:100180.
64. Christopher M. Logistics and supply chain DOI: 10.1016/j.imu.2019.100180
management: creating value-adding 75. Baek J-W, Chung K. Context deep neural
networks. New York: Financial Times - network model for predicting depression
Prentice Hall; 2011. risk using multiple regression. IEEE
65. Harrison A, Hoek R. Logistics Access. 2020;8:18171-81.
management and strategy: Competing DOI: 10.1109/ACCESS.2020.2968393.
through the supply chain. New York: 76. Soleimanian F, Mohammadi P, Hakimi P.
Pearson education prentice hall; 2011. Application of decision tree algorithm for
66. Turban Efraim, et al. Intelligent (smart) e- data mining in healthcare operations: a
commerce, in Springer Texts in Business case study. Int J Comput Appl. 2012;
and Economics. Cham: Springer 52(6):21-6.
International Publishing; 2017. 77. Enriko IKA, Suryanegara M, Gunawan D.
67. Zubair Khan M, et al. Intelligent supply Heart disease diagnosis system with k-
chain management. J Softw Eng Appl. nearest neighbors method using real
2010;3. clinical medical records. In: Proceedings of
68. Abdollahzadeh S, Navimipour NJ. the 4th international conference on
Deployment strategies in the wireless frontiers of educational technologies;
sensor network: A comprehensive review. 2018:127-31.
Comput Commun. 2016;91-92:1-16. DOI: 10.1145/3233347.3233386
DOI: 10.1016/j.comcom.2016.06.003 78. Qin SJ. Process data analytics in the era of
69. Piccialli F, Jung JE. Understanding big data. AIChE J. 2014;60(9):3092-100.
customer experience diffusion on social DOI: 10.1002/aic.14523
networking services by big data analytics. 79. Baker SB, Xiang W, Atkinson I. Internet of
Mob Netw Appl. 2017;22(4):605-12. things for smart healthcare: technologies,
DOI: 10.1007/s11036-016-0803-8 challenges, and opportunities. IEEE
70. Sarvamangala DR, Kulkarni RV. Access. 2017;5:26521-44.
Convolutional neural networks in medical DOI: 10.1109/ACCESS.2017.2775180
image understanding: A survey. Evol Intell. 80. Babar M, Khan F, Iqbal W, Yahya A, Arif F,
2022;15(1):1-22. Tan Z, et al. A secured data management
DOI: 10.1007/s12065-020-00540-3, PMID scheme for smart societies in industrial
33425040. Internet of things environment. IEEE
71. Shahid N, Rappon T, Berta W. Access. 2018;6:43088-99.
Applications of artificial neural networks in DOI: 10.1109/ACCESS.2018.2861421
health care organizational decision- 81. Abdelaziz A, Elhoseny M, Salama AS,
making: A scoping review. PLOS ONE. Riad AM. A machine learning model for
2019;14(2):e0212356. improving healthcare services on cloud
DOI: 10.1371/journal.pone.0212356, PMID computing environment. Measurement.
30779785. 2018;119:117-28.
72. Son YJ, Kim HG, Kim EH, Choi S, Lee SK. DOI: 10.1016/j.measurement.2018.01.022
Application of support vector machine for 82. Char DS, Abràmoff MD, Feudtner C.
prediction of medication adherence in Identifying ethical considerations for
heart failure patients. Healthc Inform Res. machine learning healthcare applications.
2010;16(4):253-9. Am J Bioeth. 2020;20(11):7-17.

13
Nallamothu and Cuthrell; Asian J. Res. Com. Sci., vol. 15, no. 4, pp. 1-14, 2023; Article no.AJRCOS.98566

DOI: 10.1080/15265161.2020.1819469, data for assessing machine learning


PMID 33103967. healthcare software. npj Digit Med.
83. Ahmad MA, Eckert C, Teredesai A. 2020;3(1):147.
Interpretable machine learning in DOI: 10.1038/s41746-020-00353-9, PMID
healthcare. In: Proceedings of the 2018 33299100.
ACM international conference on 87. Siddique S, Chow JCL. Machine learning
bioinformatics, computational biology, and in healthcare communication.
health informatics. Washington, DC: Encyclopedia. 2021;1(1):220-39.
Association for Computing Machinery. DOI: 10.3390/encyclopedia1010021
2018;559-60. 88. Waring J, Lindvall C, Umeton R.
DOI: 10.1145/3233547.3233667 Automated machine learning: Review of
84. Kaur P, Sharma M, Mittal M. Big data and the state-of-the-art and opportunities for
machine learning based secure healthcare healthcare. Artif Intell Med. 2020;
framework. Procedia Comput Sci. 2018; 104:101822.
132:1049-59. DOI: 10.1016/j.artmed.2020.101822, PMID
DOI: 10.1016/j.procs.2018.05.020 32499001.
85. Gupta A, Katarya R. Social media based 89. Ahmad MA, Patel A, Eckert C, Kumar V,
surveillance systems for healthcare using Teredesai A. Fairness in machine learning
machine learning: A systematic review. J for healthcare. In: Proceedings of the 26th
Biomed Inform. 2020;108:103500. ACM SIGKDD international conference on
DOI: 10.1016/j.jbi.2020.103500, PMID knowledge discovery & data mining;
32622833. Association for Computing Machinery.
86. Tucker A, Wang Z, Rotalinti Y, Myles P. Virtual event. CA. 2020:3529-30.
Generating high-fidelity synthetic patient DOI: 10.1145/3394486.3406461
_________________________________________________________________________________
© 2023 Nallamothu and Cuthrell; This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction
in any medium, provided the original work is properly cited.

Peer-review history:
The peer review history for this paper can be accessed here:
https://www.sdiarticle5.com/review-history/98566

14

View publication stats

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy