A Compendium of Various Applications of Machine Learning

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 07 | July 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1141
A Compendium of Various Applications of Machine Learning
Meena Siwach1, Suman Mann2
1Research Scholar, USICT, GGSIPU, Delhi
1Assistant Professor, Department of Information Technology, Maharaja Surajmal Institute of Technology,
meenusiwach@gmail.com
2Associate Professor, Department of Information Technology, Maharaja Surajmal Institute of Technology,
Suman mann2007@gmail.com
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Machine learning, or ML, is a branch of study
that largely focuses on computer programs that use data to
“learn”, that is, identify patterns and relations in data which
is difficult to achieve using conventional programming, and
methods to solve problems more accurately over time.
Machine Learning was largely considered a subfield of
Artificial Intelligence since its inception. It was only by
1990’s when Machine Learning started to flourish as a
separate field which caused Machine Learning to become
widely known and widely used in various technologies. This
paper aims to review various applications of Machine
Learning.
Key Words: Machine Learning, Artificial Intelligence,
Anomaly Detection, Algorithms, Deep Learning, Cancer
Prediction
1. INTRODUCTION
The growing area of data science includes machine learning
as a key element. Algorithms are trained to generate
classifications or predictions using statistical techniques,
revealing important insights in data mining operations. The
decisions made as a result of these insights influence key
growth indicators in applications and enterprises. Much of
machine learning (ML) research is inspired by weighty
problems from biology, medicine,finance,astronomy,etc[1].
The use of machine learning techniques in bioinformatics
includes genomics, proteomics, microarrays, systems
biology, evolution, and text mining. Two areas which may
benefit from the application of ML techniques in the medical
field are diagnosis and outcome prediction. This includes a
possibility for the identification of high risk for medical
emergencies such as relapse or transition into another
disease state [2]. Machine learning algorithmsareemployed
in the financial industry to spot fraud, automate trading,and
offer investors financial advising services. Much of finance
involves pattern recognition using data, where multifarious
inputs are modeled to predict outputs. For example, stock
market prediction may be based on many variables
(streaming data on stock prices, interest rates, volatilities,
etc.). Another case is in consumer banking,wherecustomers
are characterized by myriad variables to determine what
products to offer them, or to compute their probabilities of
retention [3]. In the field of astronomy, Machine Learning is
applied for classification of galaxy type, classifying different
types of stars, to predict the mass of a local group etc. In this
paper, we will delve into some of theapplicationsofMachine
Learning in detail.
Fig-1: Venn diagram showing relationship between
Artificial Intelligence, Machine Learning and Deep
Learning
2.VARIOUSAPPLICATIONSOFMACHINELEARNING
2.1 Energy Efficiency in Industry
Large amounts of data are produced more frequently than
not in the modern industrial environment, but most of it
seems to go unused by most businesses. The topic of EM in
industry received very little attention prior to the 1970s [5].
The oil crisis of the 1970s, which raised worries about
energy security and activelypromotedmore energy-efficient
technology and practices, was one of the turning moments.
Because the scientific community has been activelyinvolved
in tackling these difficulties, there is a wealth ofliteratureon
the subject covering a wide range of industrial EE-related
topics. Multiple contributions and a range of managerial
tactics have been documented for promoting EE. The
application and efficiency of energy audits in industry are
topics that are covered in several contributions. The

possibility of modeling methodologies for EE objectives is
investigated by Zhou et al. (2016) [6]. The development of
optimization-based approaches or of predictionstrategiesis
made possible by modeling. A wide variety of tools are
included in ML for the extraction of knowledge from data.
This comprises Principal Component Analysis (PCA),
Support Vector Machines (SVM), Clustering techniques,
Artificial Neural Networks (ANN), and many other methods.
A variety of procedures must be used to systematically
modify raw data to gain ever richer information because the
problem of extracting insights from data is frequentlynotan
easy one. A typical data process for insight generation is
shown in Fig-2.
Fig-2: Typical procedures for using ML technologies to
derive insight from data [6]
2.2 Approaches in Intrusion Detection System
One of the main issues of the modern day is network
security. The vulnerabilities of network securityhavegrown
in importance as a result of the internet's rapid expansion
and widespread use over the past ten years. Unauthorized
access and unexpected attacks on secured networks are
found using intrusion detection systems. Numerous studies
on the intrusion detection system have been undertaken in
recent years [7]. As a solitary classifier or single classifier,
one machine learning method or techniquecanbeappliedto
the development of an intrusion detection system. Decision
Tree, Naive Bayes, K-nearest Neighbours, Artificial Neural
Network, Support Vector Machines, and Fuzzy Logic are a
few machine learning algorithms that have been discovered
to be widely utilized single classifiers. The primary idea
underlying SVM for intrusion detection is to assume thatthe
remaining items are anomalies by just using the training
data to describe the typical class of objects,or whatisknown
as a non-attack in an intrusion detection system.
A hybrid classifier combines many machine learning
methods or methodologies to significantly increase the
performance of the intrusion detection system. Using pre-
processing methods based on clustering to weed out non-
representative training samples from the training data,
followed by using the clustering results as training samples
for pattern recognition to create a classifier. Weak learners
are classifiers that perform only marginally better than a
random classifier. Ensemble classifiers are used when
several weak learners are merged with the goal of
considerably increasing a classifier's performance.
2.3 Anomaly Detection
Since many years ago, anomalydetectionhasbeenutilized to
locate and separate aberrant components from data.
Anomalies have been found using a variety of ways.Machine
Learning (ML), which is one of the increasingly important
techniques, is crucial in this area. Finding patterns in data
that do not match expected behaviour is known as anomaly
detection [8].
In order to construct models that could detect anomalies
when used, researchers used 28 ML approaches, as
illustrated in Fig-3. Classification, ensemble, optimization,
rule system, clustering, andregressionaresixcategoriesinto
which these methodscanbe separated.Intrusiondetectionis
a persistent issue in the realm of computer security.
Anomaly detection is one of the many viable ways to
intrusion detection that has attracted a lot of interest [9].
Fig-3: Different machine learning techniques used or anomaly detection [8]

2.4 Quantitative Finance
In a variety of sectors, such as fraud detection, payment
processing, and regulation, ML is being used to enhance
function throughout the financial sector. [10] Systems that
mimic human thought processes include AI and ML. These
days, a lot of these technologies are marketed as cognitive
computing systems. Recent years have seen an increase in
the use of machine learning (ML) techniques,aswell asa rise
in interest in their financial applications. These applications
include sentiment analysis of news, trend analysis, portfolio
optimization, and risk modelling, among many other use
cases that support investment management.
With the development of computing technology, it became
possible to gather and analyse massive volumes of market
data, which led to the rise in popularity of a quantitative
approach to market analysis. This led to tremendous
advancements in our understanding of financial markets by
enabling the development and verificationof marketmodels
on a scale that was previously impractical.
Support vector machine (SVM)isa machinelearningmethod
for categorizing data. A decision function thatmaximizesthe
margin between classes is discovered by the SVM. [11] is a
noteworthy SVM-based effort in whichtheauthorssought to
create a productive technique for making large gains by
studying stock markets. The SVM model was specifically
used by the authors to pick only equities that outperformed
the market in terms of percentage return. The outcomes
supported the success of the suggested SVM; in fact, the
stocks chosen showed a total return of 208 percent overa 5-
year period.
2.5 Cancer Prediction and Prognosis
A developing trend towards personalized, predictive
medicine includes the use of computers (and machine
learning) in disease predictionandprognosis.Inmore recent
times, machine learning has been used to forecast and
prognostic cancer. In the field of cancer research, machine
learning is not new. For almost 20 years, cancer detection
and diagnosis have relied on artificial neural networks
(ANNs) and decision trees (DTs). Predictive medicine is a
growing field, and it's crucial for patients, doctors, health
economists, and policy makers as well (in implementing
large scale cancer prevention or cancer treatment policies).
The use of machine learning techniques today spans a wide
range of applications, from the detectionandclassificationof
tumors using X-ray and CRT images to the classification of
malignancies from proteomic and genomic (microarray)
studies. [12] The primary objectives of cancerprognosisand
prediction are different from those of cancer diagnosis and
detection. Three predictive foci are important in cancer
prognosis and prediction: 1) cancersusceptibilityprediction
(risk assessment); 2) cancer recurrence prediction; and 3)
cancer survival prediction. These kinds of molecular-scale
details on patients or tumors can now be easily gathered
thanks to the quick development of genomic,proteomic, and
imaging technology.
3. CONCLUSIONS
Machine Learning provides a wide variety of helpful
approaches to issues that might otherwise defy manual
resolution. It allows computers to identify patterns and
relations with minimal effort. It also improves the
performance with “Experience” that is with every execution,
the program optimizes its performance and gets more
accurate. Machine learning has advanced recentlyasa result
of the creation of new learning theories and algorithms as
well as the continual explosion in the accessibility of online
data and low-cost processing. Science, technology, and
business have all adopted data-intensive machine-learning
techniques, which has increased the use of evidence-based
judgment in numerous fields such as marketing,
manufacturing, health care, and financial modeling.Machine
Learning can be adopted to solve problemsorprovidebetter
insights for any industry or field, as clearly demonstrated in
this paper. The applications discussed in this short paper
encompass energy, security, finance and healthcare. The
real-life applications of Machine Learning are endless and
rapidly growing at a very steady pace. Today everypersonis
using Machine Learning either with his or her knowledge or
without. It is not far-fetched to say that in the near future
Machine Learning will be impacting our lives directly, even
in the aspects we cannot yet imagine.
REFERENCES
[1]: Wagstaff, K., 2012. Machine learning that matters. arXiv
preprint arXiv:1206.4656.
[2]: Sidey-Gibbons, J., Sidey-Gibbons, C. Machine learning in
medicine: a practical introduction. BMC Med Res Methodol
19, 64(2019).https://doi.org/10.1186/s12874-019-0681-4
[3]: Culkin, R. and Das, S.R., 2017. Machine learning in
finance: the case of deep learning for option pricing. Journal
of Investment Management, 15(4), pp.92-100.
[4]: Sebastiani,Fabrizio."Machinelearninginautomated text
categorization. " ACM computing surveys (CSUR) 34. 1
(2002): 1-47.
https://dl.acm.org/doi/10.1145/505282.505283
[5]: Application of machine learning tools for energy
efficiency in industry: A review by Diogo A.C.Narciso,
F.G.Martins
https://www.sciencedirect.com/science/article/pii/S23524
84719308686
[6]: Energy consumption model and energy efficiency of
machine tools: a comprehensive literaturereviewby Zhou et
al.

https://linkinghub.elsevier.com/retrieve/pii/S0959652615
006617
[7]: Application of Machine Learning Approaches in
Intrusion Detection System: A Survey by Nutan Farah Haq,
Abdur Rahman Onik, Md. Avishek Khan Hridoy, Musharrat
Rafni, Faisal Muhammad Shah, Dewan Md. Farid
https://pdfs.semanticscholar.org/bbdf/15442913c6145ce8
e9650088b8c0f8ab3c66.pdf
[8]: A. B. Nassif, M. A. Talib, Q. Nasir and F. M. Dakalbab,
"Machine Learning for Anomaly Detection: A Systematic
Review," in IEEE Access, vol. 9, pp. 78658-78700, 2021, doi:
10.1109/ACCESS.2021.3083060.
[9]: An Application of Machine Learning to Anomaly
Detection, Terran Lane and Carla E.Brodley,IN 47907-1287,
February 14, 1997.
[10]: Emerson, Sophieand Kennedy,RuairíandO'Shea,Luke
and O'Brien, John, Trends and Applications of Machine
Learning in Quantitative Finance (May 30, 2019). 8th
International Conference on Economics and Finance
Research (ICEFR 2019),
https://ssrn.com/abstract=3397005
[11]: Fan, A.; Palaniswami, M. Stock selection using support
vector machines. In Proceedings of the International Joint
Conference onNeural Networks(IJCNN’01),Washington, DC,
USA, 15–19 July 2001; Volume 3, pp. 1793–1798.
[12]: Applications of Machine Learning in Cancer Prediction
and Prognosis by Joseph A. Cruz, David S. Wishart, January1,
2006
https://doi.org/10.1177/117693510600200030
[13]: Mann, Suman, et al. "Estimation of age groups using
facial recognition features" International journal of
engineering and computer science,2018 pp 23945-23951.
[14]: 24. Hooda, S., and S. Mann. "A Focus on the
ICU’s Mortality Prediction Using a CNN-LSTM Model."
International Journal of Psychosocial Rehabilitation24,no.6
(2020): 8045-8050.
[15]: 25. Vasu Negi, Suman Mann , Vivek Chauhan, “
Devanagari Character Recognition Using Artificial Neural
Network”, International Journal of Engineering and
Technology, 2017, 2161-2167
[16]: 26. Suman Mann, Deepa Gupta, Yukti Arora,
Shivanka Priyanka Chugh, Akash Gupta, Smart Hospitals
Using Artificial Intelligence and Internet of Things for
COVID-19 Pandemic, chapter in Smart Healthcare
Monitoring Using IoT with 5G, 2021

A Compendium of Various Applications of Machine Learning

More Related Content

Similar to A Compendium of Various Applications of Machine Learning (20)

More from IRJET Journal (20)

Recently uploaded (20)

A Compendium of Various Applications of Machine Learning