Machine Learning

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Materials Today: Proceedings xxx (xxxx) xxx

Contents lists available at ScienceDirect

Materials Today: Proceedings


journal homepage: www.elsevier.com/locate/matpr

IMPACT OF MACHINE learning ON Management, healthcare AND


AGRICULTURE
Harikumar Pallathadka a,⇑, Malik Mustafa b, Domenic T. Sanchez c, Guna Sekhar Sajja d, Sanjeev Gour e,
Mohd Naved f
a
Director, Manipur International University, Imphal, Manipur, India
b
Center for foundation Studies, Gulf College
c
College of Education, Cebu Technological University, Cebu, Philippines
d
Research Scholar, University of the Cumberlands, USA
e
Career College, Bhopal
f
Department of Business Analytics, Jagannath University, Delhi-NCR, India

a r t i c l e i n f o a b s t r a c t

Article history: In the agriculture and healthcare industries, AI has been deployed to achieve better crop production, dis-
Available online xxxx ease prediction, continuous monitoring, efficient supply chain management, improved operational effi-
ciency, and reduced water waste, with the main goal of designing standard, reliable product quality
Keywords: control methods and the search for new ways of reaching and serving society while maintaining low cost.
Artificial Intelligence Machine learning and deep learning are two of the most often used AI approaches. Individuals, busi-
Machine Learning nesses, and government agencies utilize these models to anticipate and learn from data. Machine learn-
E-commerce
ing models for the complexity and diversity of data in the real-world applications are being developed at
Finance
Prediction
the moment. This article discusses machine learning and artificial intelligence applications in agriculture,
Analysis healthcare, management, and social studies. Disease Prediction, water irrigation optimization, Sales
growth, profit maximization, sales forecast, inventory management, security, fraud detection, and port-
folio management are some of the major uses.
Ó 2021 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific
committee of the International Conference on Nanoelectronics, Nanophotonics, Nanomaterials, Nano-
bioscience & Nanotechnology.

1. Introduction Agriculture is important to the country’s economy since it feeds


the whole population. It links and interacts with all of the country’s
Artificial intelligence is commonly used to describe the artificial relevant enterprises in this way. A country is considered socially
creation of human brains capable of learning natural language, and economically affluent if it has a relatively large agricultural
planning it, perceiving it, or processing it [1]. It is the theory and basis. In the majority of countries, agriculture is the primary source
development of computer systems capable of doing tasks that need of employment. Large farms usually require the hiring of extra
human intelligence, such as visual perception, speech recognition, workers to help with planting and farm animal care. The bulk of
decision-making, and language translation. Artificial intelligence is these huge farms have close processing plants where their agricul-
an IT sector that mostly works with machines that are designed to tural products are completed and developed [5].
function similarly to humans. AI, according to John McCarthy (AI’s AI has been used in agriculture and healthcare to improve crop
father), is ‘‘the scientific and technological understanding of build- production, disease prediction, continuous monitoring, efficient
ing smart computer programs in particular.” [2] supply chain management, operational efficiency, and water
Machine learning and deep learning are two of the most often waste, with the main goal of designing standard, reliable product
used AI approaches. Individuals, businesses, and government agen- quality control methods and the search for new ways to reach
cies utilize these models to anticipate and learn from data. and serve society while maintaining l Deep learning and machine
Machine learning models for the complexity and diversity of data learning are two of the most popular AI techniques. These models
in the food industry are being developed at the moment [3 4]. are used by individuals, corporations, and government agencies to
predict and learn from data. Machine learning models for the com-
⇑ Corresponding author. plexity and diversity of data in real-world applications are cur-
E-mail address: harikumar@miu.edu.in (H. Pallathadka). rently being developed. This article addresses the uses of

https://doi.org/10.1016/j.matpr.2021.07.042
2214-7853/Ó 2021 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on
Nanoelectronics, Nanophotonics, Nanomaterials, Nanobioscience & Nanotechnology.

Please cite this article as: H. Pallathadka, M. Mustafa, D.T. Sanchez et al., IMPACT OF MACHINE learning ON Management, healthcare AND AGRICULTURE,
Materials Today: Proceedings, https://doi.org/10.1016/j.matpr.2021.07.042
H. Pallathadka, M. Mustafa, D.T. Sanchez et al. Materials Today: Proceedings xxx (xxxx) xxx

machine learning and artificial intelligence in agriculture, health- uses many distance measures to find the next neighbor such as dis-
care, management, and social studies. Some of the primary applica- tance from Manhattan, Euclidean, and Minkowski etc. A kNN clas-
tions are disease prediction, water irrigation optimization, sales sifier is constructed in this work utilizing Euclidean distance
growth, profit maximization, sales forecasting, inventory manage- measurement and a k value of 3.
ment, security, fraud detection, and portfolio management [6 7]. A discriminative model is used to determine the right class
using the Support Vector Machine (SVM) classifier. The SVM model
recognizes the split between two categories accurately so that the
2. Machine learning techniques
right class of the new data point may be identified when a new
data point comes. To identify this split, a number of recognized
In general, machine learning algorithms are divided into two
techniques can be employed. A linear approach for separating the
groups: generative models and discriminative models. Various
partition from its support vectors, maximizing it. The support vec-
supervised machine learning techniques are employed in this
tors are the closest data points of each class in the training dataset.
study to analyze activity recognition. Fig. 1 depicts the relationship
The SVM [10] is similar to a hyperplane construction of equations,
between artificial intelligence and machine learning. As a result,
in which it is possible to classify a new point xi into one of the
Nave Bayes, kNN, and SVM exhibit a higher capacity to properly
classes with the function.
recognize the activities. The Nave Bayes method is a generative
In many instances, SVM classificators hesitate to split the data
model, whereas kNN and SVM (Radial Bias and Polynomial) are
input into certain classes by using a linear hyperplane. This problem
discriminative models.
is solved utilizing non-linear functions for transforming high-
The naïve classification [8] in Bayes is based on the Bayes theo-
dimensional input data. This procedure helps to distinguish the
rem and naïve independence assumptions. It is an easy classifica-
input data so that the converted space can provide a linear separable
tion technique for activity recognition. This system identifies
plane. The huge size of the space on the other hand makes it almost
human activity through the use of a vast collection of user behavior
difficult to compute the inner product of two transformed vectors.
sensor data. This classifier predicts the distribution of parameters
‘‘Kernel functions” are designed in order to deal with this issue which
based on the autonomy assumption.
are used in the feature space in the place of the internal product of
KNN [9], also k-Nearest Neighbour, classifies test data on the
two changed data vectors. By employing kernel functions correctly,
basis of the k most common known classes. The kNN classification
the computation effort for achievable tasks may be minimized.
classifier is a density classifier. This approach uses a distance mea-
For certain applications of SVM-based classification, a correct
sure such as the Euclidean distance to locate the nearest neigh-
kernel function selection is important. The intelligent choice of
bours. Two classes, ’activity eating’ and ’activity sleep’ are
kernel function allows SVM to develop learning potential. A series
present, each with the red circle and green star. There are two clas-
of kernel features were addressed in the literature. The first
sifications. Eight cases of the class ’activity-consuming’ and six
assumption by SVM is that the data set is linear and builds a hyper-
cases of the course ’activity-sleeping’ are included in the training
plane. The functionality of the kernel is then employed to decrease
data set, for a total of 14 occurrences. If the trained kNN model is
the data set error produced by SVM assumptions.
introduced to a new instance (test data), the new instance (repre-
For certain applications of SVM-based classification, a correct
sentative by a yellow square) tries to be classified into one of these
kernel function selection is important. The intelligent choice of
classes. Depending on how much votes his neighbors cast, the class
kernel function allows SVM to develop learning potential. A series
of a new instance is determined. This model takes into account the
of kernel features were addressed in the literature. The first
nearest neighbors and votes on the new instance. The model pre-
assumption by SVM is that the data set is linear and builds a hyper-
dicts that the new instance will belong to the class ‘‘Activity Sleep-
plane. The functionality of the kernel is then employed to decrease
ing” if the K value is set to 1, as it has earned more votes than its
the data set error produced by SVM assumptions.
neighbor ‘‘Activity Sleeping.” Once K is set to 3, the model predicts
that the new instance is activity-eating since seven ‘‘activity-
I. If the entire sample of the list belongs to the same category,
eating” cases occur in the K = 3 frontier. The kNN classifier model
it generates a leaf node in the decision tree to choose that
class.
II. If none of the characteristics are capable of providing infor-
mation gain, it constructs a decision node further up the tree
based on the predicted value of the class.
III. If the example originates from an earlier unaffected class, it
constructs a decision node further up the tree by the class’s
expected value.

The Alternating Decision Tree is a classification of the machine


to generalize and enhance the decision tree. This classification for
the decision tree consists of nodes of decision and forecast. Predic-
tion nodes have one number, whereas Decision nodes create a
predicate. Both the roots and the leaves can include prediction
nodes. The method of classification works by going through each
path to completion of all decision nodes and adding any prediction
nodes crossed.
A random forest [11] classifier uses the cluster method. This
method combines the choices of multiple weak classifiers to pro-
vide the best results in classification. RF combines decision trees
on the other side in order to get the greatest accurate categoriza-
tion. The tagging technique is utilized for a weak set of decision
tree classifiers, which either modifies the data samples or chooses
Fig 1. Relationship between Artificial Intelligence and Machine Learning. an arbitrary input from a feature subset.
2
H. Pallathadka, M. Mustafa, D.T. Sanchez et al. Materials Today: Proceedings xxx (xxxx) xxx

This algorithm uses a learning method which was initially sug- technology such as smart irrigation systems. However, the preci-
gested by Holland in 1975, based on natural selection ideas. The sion of such a system is dependent on soil data from that region
Genetic algorithm initially explores a sample of space with a con- as well as climate data from that region. Machine learning reduces
stant population. Each population member’s fitness is evaluated. water waste in irrigation.
Then new people are produced by selecting the ‘‘offspring” per- Agriculture is critical to the survival of humanity. Agriculture
forming individuals. The children keep their parents’ characteris- provides a living for a large portion of the world’s population. It also
tics and generate a fitter population. Two major genetic gives a great number of job opportunities for locals. Many farmers
operators, namely ‘‘Crossover” and ‘‘Mutation,” are involved in want to continue using old-fashioned farming practices that gener-
the creation of new humans. ate a low revenue. Agriculture and allied businesses are critical to the
In two parent gene structures, the crossover operator chooses a economy’s long-term development and prosperity. The key difficul-
random spot and produces two new persons by altering the ties for agricultural production are decision making, crop selection,
remaining parent parts. This generates two new people, potentially and supporting systems for enhanced crop output. Agriculture fore-
enhanced fitness, by fusioning the two elderly people. By random casting is influenced by variables such as temperature, soil fertility,
changes to an existing component, the mutation operator, on the water availability, water quality and season, crop price, and so on.
other hand, produces a new individual. The work of this operator Machine learning is useful in predicting agricultural yields based
is comparable to that of the population disturbance operator, on location, climatic data, and season. It assists farmers in cultivating
which might provide the population fresh knowledge. This opera- the best suitable crop for their agricultural property.
tor also helps prevent any stoppage during the search procedure.
3.2. Application of Machine learning and artificial intelligence
healthcare
3. Applications of Machine learning and artificial intelligence in
Agriculture, healthcare and management Heart disease is one of the most dangerous diseases for humans.
It has resulted in maximum deaths around the world. It kills many
3.1. Application of Machine learning and artificial intelligence in young people around the globe. Researchers have given a warning
agriculture that the number of heart patients will increase exponentially in
upcoming years. Applications of machine learning and artificial
Agricultural research has benefited from technological advance- intelligence in healthcare sector [17 –20] are listed as follows:
ments, particularly by incorporating industrial advances into a sus-
tainable agriculture production system. By electrifying every  Disease Prediction
farming procedure, technology has transformed farming into a  Multi disease prediction
viable business. This saves the farmer money and eliminates the  AI Assisted Surgery
middleman who buys low from farmers and sells high to end con-  Personalized Medicine
sumers. Recent applications of computational intelligence tech-  Medical Image Detection
niques provide solutions to site-specific decision modeling
problems in agricultural systems. Applications of machine learning A large number of researchers around the world are working on
and artificial intelligence in agriculture sector [12–16] are listed as heart disease patients data to anticipate or predict this disease.
follows: This is a high time to perform research on heart disease data set
available to frame a assured disease prediction model, to stop unti-
 Crop disease prediction mely death of many peoples. Machine learning algorithms are a
 Crop yield prediction great tool for classification and prediction task.
 Reducing fresh water wastage
 Soil Classification 3.3. Application of Machine learning and artificial intelligence
 Crop Monitoring and Tracking management
 Selection of Correct Crop
Application of Machine Learning and Artificial Intelligence Man-
Agriculture is important to the country’s economy since it feeds agement in management sector [21–24] is listed as follows:
the whole population. It links and interacts with all of the country’s
relevant enterprises in this way. A country is considered socially  Chatbots
and economically prosperous if it has a sufficiently large agricul-  Image search
tural basis. In the majority of countries, agriculture is the primary  Handling Customer Data
source of employment. Large farms usually require the hiring of  Recommendation Systems
extra workers to help with planting and farm animal care. The bulk  Inventory Management
of these huge farms have close processing plants where their agri-  Cybersecurity
cultural products are processed and developed.  Credit Scoring, Loan Underwriting, Portfolio management
Machine learning’s adaptability, promotion, and reduced costs
helps in assessing the complicated link between the input and out- AI has been used in the e-commerce and financial industries to
put of agricultural systems utilizing analytical approaches that are improve customer experience, efficient supply chain management,
characterized by non-linearity, time variable features, and numer- operational efficiency, and mate size, with the main goal of design-
ous unknown elements. This paper provides a review of various ing standard, reliable product quality control methods and the
machine learning-based algorithms that can be used in agriculture search for new ways of reaching and serving customers while
for tasks such as crop disease detection, intelligent irrigation, soil keeping costs low. Deep learning and machine learning are two
classification, monitoring, and tracking. of the most popular AI techniques. These models are used by indi-
According to FAO, agriculture is not only the world’s largest viduals, corporations, and government organizations to predict and
water consumer in terms of volume, but it is also a low-value, learn from data. At the moment, machine learning models for the
low-efficiency, and heavily subsidized water user. As a result, there complexity and diversity of data in the food sector are being cre-
is a need to improve agricultural irrigation efficiency by utilizing ated. This article addresses the uses of machine learning and arti-
3
H. Pallathadka, M. Mustafa, D.T. Sanchez et al. Materials Today: Proceedings xxx (xxxx) xxx

ficial intelligence in e-commerce, business management, and [5] A. Sharma, A. Jain, P. Gupta, V. Chowdary, Machine Learning Applications for
Precision Agriculture: A Comprehensive Review, IEEE Access 9 (2021) 4843–
finance. Some of the most common applications include sales
4873, https://doi.org/10.1109/Access.628763910.1109/ACCESS.2020.3048415.
growth, profit maximization, sales forecasting, inventory manage- [6] K. Shailaja, B. Seetharamulu and M. A. Jabbar, ‘‘Machine Learning in
ment, security, fraud detection, and portfolio management. Healthcare: A Review,” 2018 Second International Conference on Electronics,
Communication and Aerospace Technology (ICECA), 2018, pp. 910-914, doi:
10.1109/ICECA.2018.8474918.
4. Conclusion [7] R. Kamal, A. Karan and V. S. Arungalai, ‘‘Investigations on E-commerce Data for
Forecasting the Efficient Promotional Platform Using Supervised Machine
Learning,” 2019 4th International Conference on Recent Trends on Electronics,
AI has been used in agriculture and healthcare to improve crop Information, Communication & Technology (RTEICT), 2019, pp. 939-943, doi:
production, disease prediction, continuous monitoring, efficient 10.1109/RTEICT46194.2019.9016688.
supply chain management, operational efficiency, and water [8] A. Khajenezhad M.A. Bashiri A. Hamid Beigy distributed density estimation
algorithm and its application to naive Bayes classification, Applied Soft
waste, with the main goal of designing standard, reliable product
Computing, Volume 98, 106837 ISSN 1568–4946 2021 10.1016/j.
quality control methods and the search for new ways to reach asoc.2020.106837
and serve society while maintaining l Deep learning and machine [9] Shichao Zhang 391 2020 234 242
[10] K. Shankar S. K. Lakshmanaprabu Deepak Gupta Andino Maseleno Victor Hugo
learning are two of the most popular AI techniques. These models
C. de Albuquerque Optimal feature-based multi-kernel SVM approach for
are used by individuals, corporations, and government agencies to thyroid disease classification J Supercomput 76 2 2020 1128 1143 10.1007/
predict and learn from data. Machine learning models for the com- s11227-018-2469-4
plexity and diversity of data in real-world applications are cur- [11] B. Kalaiselvi M. Thangamani An efficient Pearson correlation based improved
random forest classification for protein structure prediction techniques,
rently being developed. This article addresses the uses of Measurement, Volume 162, 107885 ISSN 0263–2241 2020 10.1016/j.
machine learning and artificial intelligence in agriculture, health- measurement.2020.107885
care, management, and social studies. Some of the primary applica- [12] M. Schnfeld, R. Heil and L. Bittner, Big Data on a FarmSmart Farming, Big Data
in Context: T. Hoeren, B. Kolany-Raiser, Eds, (2018), 109-120.
tions are disease prediction, water irrigation optimization, sales [13] R. Varghese, S. Sharma, Affordable Smart Farming using IoT and Machine
growth, profit maximization, sales forecasting, inventory manage- Learning, in: Proceedings of the IEEE Second International Conference on
ment, security, fraud detection, and portfolio management. Intelligent Computing and Control Systems (ICICCS), 2018, pp. 645–650.
[14] R. Thombare, S. Bhosale, P. Dhemey, A. Chaudhari, Crop Yield Prediction using
Big Data Analytics, International Journal of Computer and Mathematical
CRediT authorship contribution statement Sciences 6 (11) (2017) 53–61.
[15] R. Priya R. Dharavath K. Ekaansh Crop Prediction on the Region Belts of India: A
Nave Bayes MapReduce Precision Agricultural Model Proceedings of the IEEE
Harikumar Pallathadka: Conceptualization, Methodology. International Conference on Advances in Computing, Communications and
Malik Mustafa: Data curation. Domenic T. Sanchez: Visualization, Informatics (ICACCI) 2018 99 104
Investigation. Guna Sekhar Sajja: Writing - review & editing. San- [16] V. Suryanarayana, B.S. Sathish, A. Ranganayakulu, P. Ganesan, Novel Weather
Data Analysis using Hadoop and MapReduceA Case Study, in: Proceedings of
jeev Gour: Validation. Mohd Naved: Writing - original draft, the IEEE International Conference on Advanced Computing and
Supervision. Communication Systems (ICACCS), 2019, pp. 204–207.
[17] Ravi Manne, Sneha C. Kantheti, Application of Artificial Intelligence in
Healthcare: Chances and Challenges, Current Journal of Applied Science and
Declaration of Competing Interest Technology 40 (6) (2021) 78–89, https://doi.org/10.9734/cjast/2021/
v40i631320.
The authors declare that they have no known competing finan- [18] P. Motarwar A. Duraphe G. Suganya M. Premalatha Cognitive Approach for
Heart Disease Prediction using Machine Learning 2020 International
cial interests or personal relationships that could have appeared Conference on Emerging Trends in Information Technology and Engineering
to influence the work reported in this paper. (ic-ETITE) 2020 1 5 10.1109/ic-ETITE47903.2020.242
[19] Joy Iong Zong Chen, Hengjinda P, Early Prediction of Coronary Artery Disease
(CAD) by Machine Learning Method-A Comparative Study, Journal of Artificial
References
Intelligence 3 (1) (2021) 17–33.
[20] Karunakaran P, Yasir Babiker Hamdan, Sathish, Early Prediction of Autism
[1] R.T. Adek M. Ula A Survey on The Accuracy of Machine Learning Techniques for Spectrum Disorder by Computational Approaches to fMRI Analysis with Early
Intrusion and Anomaly Detection on Public Data Sets 2020 International Learning Technique, Journal of Artificial Intelligence 2 (4) (2020) 207–216.
Conference on Data Science, Artificial Intelligence, and Business Analytics [21] Soni, N.; Sharma, E.K.; Singh, N.; Kapoor, A. Artificial Intelligence in Business:
(DATABIA) 2020 19 27 10.1109/DATABIA50434.2020.9190436 From Research and Innovation to Market Deployment. Procedia Comput. Sci.
[2] T. R. N and R. Gupta, ‘‘A Survey on Machine Learning Approaches and Its 2020, 167, 2200–2210. [CrossRef]
Techniques,” 2020 IEEE International Students’ Conference on Electrical, [22] Di Vaio, A.; Palladino, R.; Hassan, R.; Escobar, O. Artificial intelligence and
Electronics and Computer Science (SCEECS), 2020, pp. 1-6, doi: 10.1109/ business models in the sustainable development goals perspective: A
SCEECS48394.2020.190. systematic literature review. J. Bus. Res. 2020, 121, 283–314. [CrossRef]
[3] G.G. Patil R.K. Banyal Techniques of Deep Learning for Image Recognition 2019 [23] Kumar, T.; Trakru, M. The Colossal Impact of Artificial Intelligence. E-
IEEE 5th International Conference for Convergence in Technology (I2CT) 2019 Commerce: Statistics and Facts. Int. Res. J. Eng. Technol. (IRJET) 2020, 6,
1 5 10.1109/I2CT45611.2019.9033628 570–572. Available online: https://www.irjet.net/archives/V6/i5/IRJET -
[4] R. Khan, M. Abbas, R. Anjum, F. Waheed, S. Ahmed, F. Bangash, Evaluating V6I5116.pdf (accessed on 3 December 2020).
Machine Learning Techniques on Human Activity Recognition Using [24] Soni, N.; Sharma, E.K.; Singh, N.; Kapoor, A. Impact of Artificial Intelligence on
Accelerometer Data, International Conference on UK-China Emerging Businesses: From Research, Innovation, Market Deployment to Future Shifts in
Technologies (UCET) 2020 (2020) 1–6, https://doi.org/10.1109/ Business Models. arXiv 2019, arXiv:1905.02092.
UCET51115.2020.9205376.

You might also like