548 Submission
548 Submission
548 Submission
Abstract— Due to its potential to make smart selections, deep getting to know has become
extraordinarily famous in current years. Deep learning systems in use these days, which can
be in large part reliant on centralized servers, are unable to provide features like collecting
and sharing statistics, transparency, dependability, security, and dependable information
provenance. Additionally, the use of centralized statistics to teach deep getting to know
fashions makes them at risk of the single factor of failure problem. In this look at, we check
out the significance of combining deep studying and blockchain generation. We take a look at
the frame of studies on deep getting to know and blockchain integration. By creating a
subject taxonomy based on seven standards—blockchain type, deep gaining knowledge of
models, deep getting to know-precise consensus protocols, application region, services,
information types, and deployment dreams—we classify and arrange the literature. We define
the advantages and disadvantages of the maximum superior blockchain-primarily based deep
studying frameworks to give insightful discussions. The motive of this paper is to provide a
complete evaluation on blockchain enabled deep getting to know fashions for predicting
health illnesses with using Kubernetes. It also investigates the importance of using
blockchain era inside the clinical subject for ensuring the security of data. Moreover, the
efficacy of several DL strategies are validated with the aid of the use of exceptional
parameters throughout assessment. Based on the assessment, the maximum suitable and
prompt approach is chosen to increase a brand new ailment detection framework in future.
Index Terms— Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL),
Disease Prediction, Kubernetes, and Blockchain.
I. INTRODUCTION
IN the dynamic discipline of scientific studies, the system and deep mastering strategies are
quickly increasing [1]. A lot of work is presently being achieved to enhance clinical imaging
packages using these algorithms to pick out disorder diagnosis device errors that would result
in extraordinarily unclear scientific cures. Medical utility systems [2, 3] use a device and
deep mastering algorithms to help count on early infection symptoms. Convolutional
networks especially, that are part of deep learning processes, have quick evolved a manner of
their personal for reading clinical information. It makes predictions the usage of supervised or
unsupervised algorithms the use of a particular preferred dataset. For a computer system [4]
to learn as an professional that can be utilized for prediction and decision-making later,
machine and deep learning strategies are critical. Computers can also now analyze without
express programming way to the department of take a look at known as gadget studying.
Deep learning is a subset of machine studying that permits structures to analyze for certain
and understand the sector in terms of an concept hierarchy. These fields provide computer
systems intelligence in an effort to analyze information for automatic reasoning and extract
patterns based totally at the specific records [5, 6]. However, as time passes on, extra
examine information and health facility affected person data are getting to be had. There are
many unfastened assets available for getting access to the affected person's statistics, and
studies may be done to leverage exclusive computer technologies [7] for correct affected
person diagnosis and disorder detection before it becomes deadly. Nowadays, it's miles
broadly recounted that synthetic intelligence and gadget mastering are having a enormous
impact at the medical region. To discover the situation, categorize the facts, or forecast the
results, a spread of system mastering and deep learning fashions are critiques on this paper
[8]. Using system studying algorithms, a complete examination of genomic statistics may be
achieved conveniently. When function engineering and feature selection are utilized in
literature, the effects are better for both type and predictions. For the purpose of diagnosing
cardiac illness, in paper [9-11], the authors experimented with numerous gadget getting to
know and deep mastering strategies. They additionally did hyperparameter tweaking to
enhance the results' accuracy. With a excessive accuracy of seventy eight.3%, neural
networks outperformed different models including logistic regression, SVM, and ensemble
techniques. Deep gaining knowledge of is a swiftly increasing idea that features very in
addition to the human mind. It has several representations of the records and correctly
handles the specificity hassle. The challenge of medical diagnosis use deep getting to know in
an expansion of methods. Numerous studies [12, 13] show that deep mastering strategies
outperform preceding methods, reduce type mistakes charges, and are greater noise-resistant.
It has the capacity to handle sizeable amounts of data and can effortlessly decipher hard
troubles. Recently, a number of medical prognoses have incorporated a ramification of device
getting to know techniques, bio-stimulated computing procedures, and deep learning
techniques [14].
The e-Health information produced by way of medical sensors and a patient's past
scientific records have good sized capability for use in modern medicinal drug for tailor-
made treatments and in information analytics for early disorder identity. These benefits
commonly want the desired healthcare stakeholders and structures to save, percentage, or
process the e-Health statistics. The acquisition, protection, distribution, and analysis of
patient medical records (PMRs) raises questions about ability statistics misuse. The PMRs are
regularly collected and saved in the company's personal database, away from the patient's
control [15, 16]. This implies that the patient's privacy or private options can be violated via
the sharing or manipulation of the information. Because in their concerns over the
repercussions of capacity facts exploitation, sufferers and different stakeholders are reluctant
to accept facts-primarily based healthcare delivery, which inhibits the full attention of e-
benefits. In e-Health, consent management [17] is the manner that upholds the affected
person's right to determine who has access to their records, how long they are able to have it
for, and the way it'll be used. This means that a affected person should be able to authorize
get entry to to any or all in their PMRs for a provider or any business enterprise, and that they
should additionally be capable of get rid of the authorization at any time. The PMR
operations ought to additionally be obvious and auditable. In different words, the patient
should be able to check who has visible their information within the past and spot who has
access to it currently. Blockchain is a allotted ledger technology [18], has currently come into
life and has the potential to clear up the troubles with consent management in e-Health.
Blockchain is the generation that powers the Bitcoin cryptocurrency, although its makes use
of pass past those of cryptocurrencies. Its essential residences—decentralization, openness,
data integrity, and traceability drawn from cryptosystems [19, 20] and can be used to offer
the best shape for permission control for the handling of PMRs in e-Health. In widespread,
research in this location is still in its early levels, and a number of the counseled frameworks
lack prototype implementations. This paper pursuits to give a comprehensive overview on
blockchain enabled deep getting to know techniques used for detecting disease with the assist
of kubernetes.
Typically, Artificial Intelligence (AI) and Machine Learning (ML) [21], which are
hastily getting used in lots of elements of the ailment prediction systems, have the ability to
revolutionize the drug discovery procedure. Medication screening, clinical photo analysis the
usage of high-content material microscopy, public fitness evaluation, and new genetic
molecule manufacturing are a few examples of applications using AI/ML. Recently, the Deep
Learning (DL) techniques have come to light as a technique that, frequently, has been shown
to be able to making predictions that are greater correct in applications connected to the life
sciences. Kubernetes answers [22] are employed to automate the deployment and control of
numerous such containerized programs in manufacturing. The machine this is presently most
in use is Kubernetes, which Google created and made open-source. An increasingly famous
opportunity to a non-public cloud platform is a Kubernetes cluster strolling on nearby
hardware, which gives scientists scalable and simple get entry to to CPUs and GPUs.
Moreover, it's far regularly run on top of digital machines (VMs) in the cloud, and all most
important cloud companies offer on-demand Kubernetes-based box orchestration services.
The computerized scaling tools called horizontal and vertical pod autoscaling are available in
Kubernetes [23, 24], the field orchestrator that has currently overtaken different cloud
managers in popularity. Since it can best now scale by terminating and restarting the
application Pods, the Kubernetes machine's vertical scaling functionality has been in
experimental checking out for years. Although Kubernetes' vertical pod has proof-of-concept
extensions that permit vertical scalability without interruptions, this capability remains now
not gift in the middle of Kubernetes. The smallest deployment unit in Kubernetes [25], a pod
usually consists of a few closely linked containers. Although custom and even external
metrics are feasible, with the aid of default horizontal pod considers the metric of CPU usage
and automatically scales Pods in and out based totally on threshold-based judgments.
The supervised getting to know set of rules provides a computer system with a education
set of eventualities with appropriate objectives. Using this practice statistics, the algorithm
responds nicely to the capability inputs. Regression and class are the subcategories of
supervised mastering. By using class techniques, the inputs are divided into many
instructions, and the skilled machine ought to produce actions to assign concealed inputs to
those lessons. The class utilized in unsolicited mail purging is the categorization of facts into
wholesome or sickness affected. Regression [26] is a supervised technique whilst the
outcomes are non-stop instead of discrete. In evaluation to classification predictions, which
use accuracy as a performance metric, regression predictions are assessed the use of root
mean squared errors (RMSE). The gadget will decide whether to educate using a dataset or
no longer on its very own. The system isn't supplied any labelling that can be used to make
predictions. With the aid of feature gaining knowledge of from the furnished information,
unsupervised learning may be used to uncover the underlying pattern. Clustering is a
technique of unsupervised gaining knowledge of that divides inputs into businesses. These
clusters have now not been determined formerly, and it creates groupings based totally on
similarity. Semi-supervised learning assumes that the device incorporates handiest a part of
the schooling statistics. This kind of training is used to educated statistics that can
consciousness on a few findings which might be missing. On untagged facts, this kind of
method is hired for learning selection. The semi-supervised mastering algorithm has the
characteristics of both the unsupervised-mastering and supervised mastering algorithms due
to the fact it's far taught on each labelled and unlabeled records [27]. Fig 1 depicts the overall
framework with the emerging technologies currently used inside the healthcare programs.
In an active mastering state of affairs, the machine simplest gets training tags for a sure set of
events. It is applied to enhance the optimality of the substances to accumulate tags for the
goal inclusive of purposeful areas and budgets. It is normally employed inside the biological
region to learn about biological organisms and to forecast their possibilities of surviving and
producing offspring. DL is the advanced stage of gadget learning, wherein getting to know
and statistics prediction are typically done by neural networks. It is a collection of numerous
algorithms. These are hired inside the layout of difficult wellknown structures capable of
making predictions approximately any problem. It employs a deep graph with multiple
processing layers that include both linear and non-linear transformation. Diagnostic
evaluation of sickness is a difficult challenge in today's scientific fields. Understanding the
ideal scientific inspection and evaluation required to diagnose sufferers is a critical mission
[28]. Data about scientific evaluations, patient statements, treatments, prescriptions, and
different topics are produced in great quantities by way of the healthcare enterprise. The
fundamental trouble is that these reports' exceptional creates the affect of association due to
the fact the statistics changed into improperly controlled. To extract and analyze these
scientific records elegantly and correctly, data enhancements are required. There are diverse
system getting to know techniques that can be employed with unique classifiers to distribute
statistics based on those features.
A deep neural community is an advanced shape of a neural community that has severa
hidden layers in among the enter and output layers. For forecasting consequences and
figuring out relationships and patterns in the facts set, neural networks are built. Here, the
outcome is determined using diverse mastering strategies. Similar to a human neuron, a
neural network's components, or nodes, are related to each other.
The inter-unit connection power impacts how as it should be the output is produced. Multiple
neurons may be determined in each of the deep neural community's many hidden layers.
Activation feature is critical to deep neural networks. In order to attain the very ideal non-
linearity property that permits activation capabilities to approximate any characteristic,
activation capabilities must incorporate complicated mapping capabilities which are non-
linear.
The unbounded linearly weighted mixture from neurons should be crushed by means of
activation functions as well. Additionally, it's vital to save you massive values from gathering
on the top of the processing hierarchy.
Life scientists' ability to extract statistics from pix is converting because of deep learning
[29]. These techniques allow analyses that had been formerly impractical and are extra
accurate than conventional strategies. Deep getting to know strategies are an increasing
number of being used to investigate massive imaging datasets as their capability grows.
Large-scale facts analysis is considerably hampered by using the computational necessities of
deep getting to know. Many industries, together with laptop vision, speech popularity,
biological signal analysis, etc., have adopted synthetic intelligence technology, specifically
deep gaining knowledge of. The extracted records is classified by way of the classifier, and
predictions are then made based totally on this categorized facts. ML/DL test follows these
steps. The simple subcategories of device learning include supervised, semi-supervised,
unsupervised, reinforcement, and lively getting to know algorithms. Furthermore, deep
gaining knowledge of techniques use neural networks to classify statistics and make
predictions with more accuracy [30]. They are basically an improved version of system
mastering algorithms.
A. Blockchain Technology
The development of information and communication era (ICT) has ushered in a new age
in which everything is to be had at the clicking of a button [31]. Every location of human life,
along with agriculture, clever towns, business automation, smart houses, healthcare, and
many others., relies upon on those new technology and control structures. The maximum
tremendous of these applications is healthcare, that is certainly one of a person's simple
wishes. The idea of the use of blockchain to create a public allotted ledger for all Bitcoin
transactions turned into first placed forth. Following that, a number of research projects
recognition on vital components of the Blockchain generation itself, together with
performance development, resolving double-spending intrusions, and developing effective
and allotted consensus procedures. Due to the Blockchain's capability for securely obtaining
personal fitness records, it's been hired in numerous research in latest years. Few studies are
undertaken on deployment; most studies efforts are focused on illustrating the blessings of
Blockchain-primarily based smart healthcare structures and offering architectures. Some
writings, like those stated, awareness on the administration of high-quality-grained get entry
to to user information received via IoT. Blockchain is a sort of database that differs from
conventional databases like relational databases in that it is seen as a public ledger [32]. As
an alternative, blockchain uses virtual signatures to chain facts stored in blocks together via a
allotted community. New blocks are surpassed via as new information are accrued.
Persistence, secrecy, traceability, and democratization are a number of the main traits of a
blockchain. Blockchain has recently been utilized in contemporary technology like IoT-based
public and protection offerings. Additionally, organizations use blockchain to attract
customers with its tremendous reliability and protection. Additionally, because blockchain
operates in a dispensed setting, the unmarried factor of failure hassle may be unnoticed. Fig 3
shows the blockchain enabled records safety model used in healthcare packages.
Fig three. Blockchain enabled statistics safety
Blockchain saves records in a manner that makes it very tough for hackers to alter,
corrupt, or erase facts. It is a decentralized generation by using layout that uses peer-to-peer
(P2P) architecture to store and execute transactions and records. It is made up of numerous
nodes that validate and store the transactions as blocks. A set of transactions are stored in
every block of the chain, and it is confident that the blocks that already exist are efficaciously
related to the newly fashioned blocks to construct the chain of blocks. The newly introduced
block is disseminated to all involved nodes to assure facts consistency after being brought to
the chain keep by means of a miners. The blockchain transactions are tested and authorized
by means of the miner nodes because of the decentralized consensus mechanism. Examples
of consensus protocols used by many blockchain structures include Proof-of-paintings and
Proof-of-Stake, that could guard the blockchain from any either internal or external record
cyberattacks. Another critical element of blockchain era is sensible contracts, which
characteristic like an digital code and most effective execute while certain conditions are
happy. The goal of clever contracts is to lessen each chance and rate.
Immutability of facts, smart contracts, mutual authentication, and scalability are a number of
the essential qualities and factors of blockchain era that assist to growth commercial
enterprise overall performance. EHRs include patient facts from all medical doctors,
hospitals, clinics, labs, and different healthcare providers.
The PHR, then again, is a patient-centric system that is controlled by using sufferers in a at
ease setting. It can make certain that sufferers are in control of the information, that it's miles
managed through patients, and that it may be shared with different users even as adhering to
the affected person's consent management policy. Smart contracts that execute on their own
are used to use the consent control policies. But if appropriate methods aren't used to manage
huge amounts of healthcare records, blockchain era becomes high priced. The guidelines and
linkers can be quite beneficial in decreasing the statistics quantity as a way to make use of
blockchain in lots of smart healthcare. Decentralized storage systems also can properly shop
big quantities of facts at the same time as avoiding single-point-of-failure-related problems.
Medical information may be recognized, analyzed, expected, and categorized the use of
AI and ML [33]; at some stage in the past ten years, numerous AI algorithms have been
efficiently hired to manner this information. Traditional Machine Learning algorithms,
however, would carry out considerably less effectively as the statistics quantity will increase
considering that they lack the ability to analyze large amounts of facts. A multi-layered
artificial neural network (ANN) has one or extra artificial neurons in every layer. There are
one or more inputs for every neuron. The first step is to multiply each enter via a community
weight (community parameter), that is commonly initialized at random [34]. Determine the
whole weighted inputs and deviation values for every neuron, upload them collectively, and
then enter the result into the activation function (nonlinear variant function). The key thing of
NN is the activation characteristic.
The DNN defines the input and output thru a complex layer composition and fashions the
operation of the neural network (NN) in the human mind. According neurons and nonlinear
operations are covered in each layer composition (activation capabilities). Deep gaining
knowledge of has the advantage over traditional system learning in that it includes severa
hidden layers and might examine from the authentic records. It can take care of big quantities
of information, study abstract statistics depending on input, and attain excellent accuracy and
overall performance. As a end result, several teachers have applied it to the world of
healthcare. The community gains non-linearity as a result, giving it the potential to examine
extra complex capabilities. Neuronal output is the result of the activation feature, and the
neuronal output from one layer of neurons serves as the input for the subsequent layer of
neurons. The loss characteristic is used to decide whether or not the network weight is most
efficient at some stage in the iterative training manner, for the duration of which the whole
community will discover the perfect weight distribution. AI and gadget getting to know are
both covered in deep learning [35]. The most essential forms of data, consisting of snap shots,
texts, and audio indicators, are represented in latent space by using a version the use of
cutting-edge deep mastering techniques. The hyperlink among deep mastering, AI, and
system gaining knowledge of is depicted in Fig four. Deep mastering permits hardware to
execute a selection of packages with accuracy this is similar to or occasionally even superior
to that of human beings in a few situations. Machine getting to know algorithms can research
multinational statistics sequences with using data illustration techniques.
To make the proper conclusions utilizing time series facts, facts pleasant is consequently
crucial, especially for system learning algorithms. More specific, regardless of how
complicated the approach is, the version will no longer work nicely on information that has
insufficient descriptions. Consequently, function engineering is taken into account
considering, by using using the sets of capabilities from the uncooked data, it may help in the
reconstruction of records. When using a deep gaining knowledge of method, the fashions
routinely separate the excessive-level latent area traits from the uncooked facts for the reason
that they are smart. Multiple layers are a function of deep gaining knowledge of models,
wherein the higher layers extract greater fundamental attributes from the input information,
whereas lower degree layers are in duty of gathering backside level functions. Accuracy and
protection in DL are impacted by way of the number of levels. Achieving the excellent
performance for the precise application relies upon closely on tuning the hyper-parameters of
the deep getting to know version. The choice of community topology on which the version is
deployed since, it's far a critical factor to keep in mind in phrases of scalability. Because the
model is discovered completely at the server cease and every node will obtain the trained
version instance to do its supposed mission, a deep gaining knowledge of version instantiated
inside a patron-server structure, as an instance, offers a scalable answer. By the use of remote
tracking and preventing patients from relapsing into the want for instant treatment, predictive
analytics is demonstrating its price no longer handiest inside the hospital setting but also at
home. At every stage of a patient's remedy, predictive analytics assist with diagnostic,
evaluation, and remedy. Remote tracking and preventing patients from relapsing into the need
for acute care are two approaches that predictive analytics is demonstrating its price outside
of the hospital setting. Every degree of a patient's remedy benefits from the usage of
predictive analytics for diagnosis, diagnosis, and therapy.
As a end result of extensive look at on RNNs and its application to series mastering, long
quick-time period reminiscence (LSTM) has been provided as a technique to the trouble of
lengthy-term dependency. LSTMs function a more sophisticated model of RNNs, changing
the recurrent gadgets with hidden layer devices of the reminiscence cells to capture lengthy-
variety interdependence and nonlinear dynamics. Long-time period temporal correlation
mastering is made less difficult via the reminiscence cells' ability to save and output records.
Self-connections that maintain the community's temporal nation and are controlled by way of
the input gate, output gate, and overlook gate are included in this. The mechanism of "gating"
establishes the cause of every memory cell in LSTMs. The LSTM modifies its cellular
kingdom when the gate is opened. The movement of inputs and outputs from reminiscence
cells into the rest of the network is managed by way of enter and output gates. The memory
cell additionally received a overlook gate, which allows excessive-weighted output statistics
to be transferred from one cell to every other. The excessive activation degree of the input
unit determines whether or not or no longer the information can be stored in reminiscence; if
it's miles excessive, the facts may be stored within the reminiscence cellular. Additionally, a
strongly excited enter unit will send statistics to the following neuron. As an opportunity,
reminiscence cells are used to keep incredibly weighted input statistics.
IV. DISCUSSION
For modelling parameters to be exchanged between regions of various places, which include
clever cities, a distributed network that is primarily based on the hyperledger cloth is
required. In truth, the blockchain network is installation through the hyperledger cloth using
bins that stand in for companies. In fashionable, Kubernetes has the blessings of quick
coordinating pods housed on numerous nodes, and it has the ability to scale based at the wide
variety of nodes wished by using the packages. The use of Kubernetes cluster in modern
cloud environments has attracted some of researchers and builders. To ensure that the ledger
was seen to all collaborating blockchain participants, the entire agencies and kubernetes
machines were related to a single channel. As a result, all areas concerned inside the
federated mastering methods could see transactions citing the modelling parameters of
diverse areas. The smart healthcare machine has obtained growing interest as a result of the
latest transformation in scientific infrastructure. An modern idea referred to as "clever
healthcare" refers to a set of suggestions that combine management, prevention, analysis, and
remedy. Smart medical systems can hyperlink and exchange records each time, anywhere, in
contrast to traditional clinical structures. Smart healthcare differs from conventional hospital
treatment in that it emphasizes information accessibility, responsiveness, and potential
reasons. By using transportable cellular devices and a wi-fi community, clinical experts can
constantly experience, system, and compare essential scientific occurrences. Every patient's
case facts is with no trouble to be had to medical doctors at any time, letting them straight
away assemble a prognosis and remedy method.
V. CONCLUSION
REFERENCES
[2] H. Hasanova, M. Tufail, U.-J. Baek, J.-T. Park, and M.-S. Kim, "A novel blockchain-
enabled coronary heart disease prediction mechanism using machine gaining knowledge of,"
Computers and Electrical Engineering, vol. One zero one, p. 108086, 2022.
[5] N. LaPierre, C. J.-T. Ju, G. Zhou, and W. Wang, "MetaPheno: A crucial evaluation of
deep mastering and system getting to know in metagenome-based ailment prediction,"
Methods, vol. 166, pp. 74-eighty two, 2019.
[6] A. K. Sahoo, C. Pradhan, and H. Das, "Performance assessment of various machine
learning methods and deep-gaining knowledge of primarily based convolutional neural
network for fitness decision making," in Nature stimulated computing for information
technological know-how, ed: Springer, 2020, pp. 201-212.
[9] M. Wang, H. Zhang, H. Wu, G. Li, and K. Gai, "Blockchain-based totally Secure
Medical Data Management and Disease Prediction," in Proceedings of the Fourth ACM
International Symposium on Blockchain and Secure Critical Infrastructure, 2022, pp. 71-
eighty two.
[10] M. Kim, J. Yun, Y. Cho, K. Shin, R. Jang, H.-j. Bae, et al., "Deep getting to know in
medical imaging," Neurospine, vol. 16, p. 657, 2019.
[12] J. Latif, C. Xiao, A. Imran, and S. Tu, "Medical imaging the usage of system
mastering and deep gaining knowledge of algorithms: a evaluation," in 2019 2d International
conference on computing, arithmetic and engineering technology (iCoMET), 2019, pp. 1-
five.
[14] M. A. Khan, "An IoT framework for coronary heart sickness prediction based totally
on MDCNN classifier," IEEE Access, vol. Eight, pp. 34717-34727, 2020.
[15] A. Khamparia, G. Saini, D. Gupta, A. Khanna, S. Tiwari, and V. H. C. De
Albuquerque, "Seasonal crops ailment prediction and class using deep convolutional encoder
community," Circuits, Systems, and Signal Processing, vol. 39, pp. 818-836, 2020.
[16] B. Lei, M. Yang, P. Yang, F. Zhou, W. Hou, W. Zou, et al., "Deep and joint studying
of longitudinal statistics for Alzheimer's sickness prediction," Pattern Recognition, vol. 102,
p. 107247, 2020.
[18] M. Senthilmurugan and R. Chinnaiyan, "IoT and gadget learning based peer to peer
platform for crop growth and ailment monitoring machine the use of blockchain," in 2021
International Conference on Computer Communication and Informatics (ICCCI), 2021, pp.
1-five.
[19] S. Lahmiri and S. Bekiros, "Cryptocurrency forecasting with deep mastering chaotic
neural networks," Chaos, Solitons & Fractals, vol. 118, pp. 35-forty, 2019.
[20] P. V. Rane and S. N. Dhage, "Systematic erudition of bitcoin fee prediction using
device mastering techniques," in 2019 5th International Conference on Advanced Computing
& Communication Systems (ICACCS), 2019, pp. 594-598.
[24] D. T. Jose, J. Holme, A. Chakravorty, and C. Rong, "Integrating massive records and
blockchain to manage power clever grid-TOTEM framework," Blockchain: Research and
Applications, p. 100081, 2022.
[25] P. Singh and P. Sharma, "Survey of Predictive Autoscaling and Security of Cloud
Resources Using Artificial Neural Networks," in Proceedings of International Conference on
Network Security and Blockchain Technology: ICNSBT 2021, 2022, pp. A hundred and
seventy-a hundred and eighty.
[26] K. B. Raju, S. Dara, A. Vidyarthi, V. M. Gupta, and B. Khan, "Smart Heart Disease
Prediction System with IoT and Fog Computing Sectors Enabled by way of Cascaded Deep
Learning Model," Computational Intelligence and Neuroscience, vol. 2022, 2022.
[29] F. Sun, J. Sun, and Q. Zhao, "A deep getting to know technique for predicting
metabolite–ailment associations thru graph neural community," Briefings in Bioinformatics,
vol. 23, p. Bbac266, 2022.
[30] S. El-Sappagh, H. Saleh, F. Ali, E. Amer, and T. Abuhmed, "Two-level deep learning
model for Alzheimer’s disease detection and prediction of the slight cognitive impairment
time," Neural Computing and Applications, pp. 1-23, 2022.
[33] Y. Cong, M. Shintani, F. Imanari, N. Osada, and T. Endo, "A New Approach to Drug
Repurposing with Two-Stage Prediction, Machine Learning, and Unsupervised Clustering of
Gene Expression," OMICS: A Journal of Integrative Biology, 2022.