MACHINE LEARNING:A REVIEW

Machine Learning :A Review

Diploma in Computer Engineering
1
Dr.Panjabrao Deshmukh Polytechnic College,
Amravati
Bhumika Thakare 1 Savi Dhoble 1
[email protected] [email protected]

Abstract

One of today’s most rapidly growing technical fields Machine learning addresses the question
of how to build computers that improve automatically through experience. Based on the
importance and potentiality of “Machine Learning” to analyze the data mentioned below, in
this paper, we provide a comprehensive view on various types of machine learning
algorithms that can be applied to enhance the intelligence and the capabilities of an application.
Machine learning is technique that improves system performance by learning from experiences
via computational method. This paper analyses the basic classification of machine learning,
including supervised learning, unsupervised learning, and reinforcement learning. Thus, this
study’s key contribution is explaining the principles of different machine learning techniques
and their applicability in various real-world application domains, such as
cybersecurity systems, smart cities, healthcare, e-commerce, agriculture, and many more. We
also highlight the challenges and potential research directions based on our study. Overall, this
paper aims to serve as a reference from the technical point of view.

Keywords
Supervised learning, learning method, unsupervised learning,
cybersecurity systems,sparisity,bayesian

i
Dr.Panjabrao Deshmukh Polytechnic College
 MACHINE LEARNING:A REVIEW

1. Introduction

Machine Learning suddenly became one of the most critical domains of Computer Science[1]
and just about anything related to Artificial Intelligence. Machine technology based on
computer technology incorporates multidisciplinary theoretical knowledge, such as statistics
and algorithm complexity, which further strengthens the functional attributes of artificial
intelligence. Machine learning involves a broad range of techniques for learning from data.
Many techniques learn a function f : R d → R that inputs attributes or features about an item,
and produces an output prediction about that item. Machine learning (ML) is a type of artificial
intelligence (AI) that allows software applications to become more accurate at predicting
outcomes without being explicitly programmed to do so. Machine learning algorithms use
historical data as input to predict new output values. Machine learning (ML) is a field of inquiry
devoted to understanding and building methods that 'learn', that is, methods that leverage [88]
data to improve performance on some set of tasks. Machine learning algorithms build a model
based on sample data, known as training data, in order to make predictions or decisions without
being explicitly programmed to do so. A subset of machine learning is closely related
to computational statistics, which focuses on making predictions using computers, but not all
machine learning is statistical learning[4]. The study of mathematical optimization delivers
methods, theory and application domains to the field of machine learning. . Data mining is a
related field of study, focusing on exploratory data analysis through unsupervised
learning.[5][6] Some implementations of machine learning use data and neural networks in a
way that mimics the working of a biological brain.[7][8] In its application across business
problems, machine learning is also referred to as predictive analytics.

2 . Classification of Machine Learning

I. Supervised Learning
In the process of machine learning, supervised learning belongs to a relatively basic
learning method. This learning method refers to the establishment of corresponding
learning goals by people before learning. During the initial training of the machine, the
machine relies on information technology to learn the needs of learning . Compared with
other learning methods, supervised learning can fully stimulate the generalized learning
potential of the machine itself. After completing the system learning, it can help people to
solve some classification or regression problems, which is highly systematic . Supervised

ii
Dr.Panjabrao Deshmukh Polytechnic College
 MACHINE LEARNING:A REVIEW

learning can be separated into two types of problems when data mining—classification and
regression:

i. Classification uses an algorithm to accurately assign test data into specific categories.
It recognizes specific entities within the dataset and attempts to draw some
conclusions on how those entities should be labeled or defined. Common
classification algorithms are linear classifiers, support vector machines (SVM),
decision trees, k-nearest neighbor, and random forest, which are described in more
detail below.
ii. Regression is used to understand the relationship between dependent and independent
variables. It is commonly used to make projections, such as for sales revenue for a
given business. Linear regression, logistical regression, and polynomial regression
are popular regression algorithms.

Figure I.Visual Representaion of linear regression

II. Unsupervised Learning

Corresponding to supervised learning is unsupervised learning. The so-called unsupervised
learning means that the machine does not mark the content in a certain direction during the
entire learning. process, but rely on the machine itself to complete the analysis of data
information. In practice, the operation method is to let the machine learn the basic concepts
and content, and then give the machine enough freedom to complete a series of content
learning, including concepts and content similar to the basic principles, such as tree roots.
In general, the continuous improvement of learning in stages has increased the breadth of
machine learning content. At present, unsupervised learning includes algorithms such as

iii
Dr.Panjabrao Deshmukh Polytechnic College
 MACHINE LEARNING:A REVIEW

deep belief networks and autoencoders. Such situations are conducive to the solution of
clustering problems and have good applications in the development of many industries.

Figure II .Types pf unsupervised learning

III. Reinforcement Learning

In addition to supervised learning and unsupervised learning, there are also application
methods of reinforcement learning in machine learning. The so-called reinforcement
learning is the systematic learning of a certain content. In the specific application process,
the data collected in the previous period will be used. It organizes and processes the
feedback information of a certain part to form a closed loop of data processing. On the
whole, reinforcement learning is a type of learning method that expands data collection
based on statistics and dynamic learning. Such methods are mainly used to solve the control
problem of robots. Its representative learning methods include Q-learning algorithm and
Temporal difference learning algorithm.

3. Research on Machine Learning Development

I. Theoretical System Continues to Mature

iv
Dr.Panjabrao Deshmukh Polytechnic College
 MACHINE LEARNING:A REVIEW

In the future development process, the mechanical theory system will also be further
optimized, and its content branches and coverage will also be expanded. In the initial
formulation process of machine learning content, its content is mainly applicable to some
automation industries, and the content of the entire theoretical system has not been completely
sound.

II. Autonomous Learning Ability is Further Improved

At present, many enterprises in China have realized the development model of automation, and
intelligence is the focus of the next stage of development. In the context of the rapid
development of Internet technology, the autonomous learning ability of machines will be
further strengthened. Whether it is supervised learning or unsupervised learning, the autonomy
that machine learning can master will continue to increase.

III. Promotion of Personalized Customization Services

With the continuous improvement of socio-economic level, people's requirements for
personalized applications are also constantly rising, which is also one of the important
development directions of machine learning in the future. With the continuous improvement
of the intelligent level of mechanical learning, different application modules can be set up
according to the actual needs of users.

4. Conclusion
In summary, machine learning is still in its infancy, and it mainly relies on supervised learning,
and does not fully overcome weak artificial intelligence. Relevant personnel need to constantly
improve the theoretical foundation and practice of machine learning. In the corresponding
scientific field and the development of computer technology, we should provide a good
environment for machine learning, and the development prospect of machine learning is very
broad. In addition, it is also necessary to actively learn from the experiences and lessons of
developed countries, set up machine algorithms suitable for the development of domestic
enterprises, and provide technical support for the economic development of the industry.

v
Dr.Panjabrao Deshmukh Polytechnic College
 MACHINE LEARNING:A REVIEW

References

[i] M. D. Zeiler and R. Fergus. Stochastic pooling for regularization of deep convolutional
neural networks. CoRR, abs/1301.3557, 2013.

[ii] Research on Machine Learning and Its Algorithms and Development To cite this
article: Wei Jin 2020 J. Phys.: Conf. Ser. 1544 012003 View the article online for
updates and enhancements.

[iii] Evaluating Machine Learning Algorithms for Fake News Detection"

by Shlok Gilda in the 2017 IEEE 15th Student Conference on Research and
Development (SCOReD), December 2017, pp.110-115

[iv] Y. Bengio, "Learning deep architectures for AI", Found. Trends Mach. Learn., vol. 2,
no. 1, pp. 1-127, 2009.

[v] M. Ranzato, F. J. Huang, Y. Boureau and Y. LeCun, "Unsupervised learning of

invariant feature hierarchies with applications to object recognition", Proc. Computer
Vision and Pattern Recognition Conf., 2007.

[vi] S. Wang and C. D. Manning. Fast dropout training. In Proceedings of the 30th
International Conference on Machine Learning, pages 118–126. ACM, 2013.

[vii] S. J. Nowlan and G. E. Hinton. Simplifying neural networks by soft weight-sharing.

Neural Computation, 4(4), 1992.

[viii] M. Chen, Z. Xu, K. Weinberger, and F. Sha. Marginalized denoising autoencoders for
domain adaptation. In Proceedings of the 29th International Conference on Machine
Learning, pages 767–774. ACM, 201

vi
Dr.Panjabrao Deshmukh Polytechnic College