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

net/publication/320609700

A Review on Machine Learning Algorithms, Tasks and Applications

Article · October 2017

CITATIONS READS
43 14,957

2 authors, including:

Neeraj Kumar
Chitkara University
47 PUBLICATIONS   166 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Handwritten Gurmukhi Character Recognition System. View project

Laptop security system View project

All content following this page was uploaded by Neeraj Kumar on 25 October 2017.

The user has requested enhancement of the downloaded file.

 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 6, Issue 10, October 2017, ISSN: 2278 – 1323

A Review on Machine Learning Algorithms,

Tasks and Applications
Diksha Sharma1, Neeraj Kumar2

learning tasks and problems and various

machine learning algorithms.
ABSTRACT: Machine learning is a field of
computer science which gives computers an Keywords: Machine learning, supervised
ability to learn without being explicitly learning, unsupervised learning,
programmed. Machine learning is used in a classification
variety of computational tasks where
1. INTRODUCTION
designing and programming explicit
Machine learning is a branch of artificial
algorithms with good performance is not
intelligence that allows computer systems to
easy. Applications include email filtering,
learn directly from examples, data, and
recognition of network intruders or
experience. Through enabling computers to
malicious insiders working towards a data
perform specific tasks intelligently, machine
breach. One of the foundation objectives of
learning systems can carry out complex
machine learning is to train computers to
processes by learning from data, rather than
utilize data to solve a specified problem. A
following pre-programmed rules. Increasing
good number of applications of machine
data accessibility has endorsed machine
learning like classifier training on email
learning systems to be trained on a bulky
messages in order to differentiate between
pool of examples, while growing computer
spam and non-spam messages, fraud
processing power has supported the critical
detection etc. In this article we will focus on
capabilities of these systems. Within the
basics of machine learning, machine
field itself there have also been algorithmic
advances, which have given machine
learning better power. As a outcome of these
advances, systems which performed at

1548
All Rights Reserved © 2017 IJARCET
 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 6, Issue 10, October 2017, ISSN: 2278 – 1323

noticeably below-human levels can now go  Supervised learning

better than humans at some definite tasks.  Unsupervised learning
Many people now cooperate with systems  Reinforcement learning
based on machine learning each day, for
Supervised Learning: It is the machine
example in image recognition systems.
learning task of inferring a function from
Now-a-days the concept of machine learning
labeled training data. The training data
is used in many applications and is a core
consists of a set of training examples. A
concept for intelligent systems [1][3] .As the
supervised learning algorithm analyzes the
field develops further, machine learning
training data and produces an inferred
shows promise of supporting potentially
function that can be utilized for mapping
transformative advances in a range of areas,
fresh examples. To work out on a given
and the social and economic opportunities
problem of supervised learning, one has to
which follow are significant. In healthcare,
carry out the following steps:
machine learning is creating systems that
can assist doctors give more correct or (i) Decide the kind of training examples.
efficient diagnosis for definite conditions. The user should decide what kind of data is
For public services it has the potential to to be used as a training set.
target support more effectively to those in
(ii) Collect a training set. The training set
need, or to tailor services to users. Machine
needs to be envoy of the real-world use of
learning is helping to make sense of the
the function. Thus, a set of input objects is
gigantic quantity of data accessible to
collected and corresponding outputs are also
researchers today, offering new insights into
collected.
biology, physics & medicine.
(iii) Decide the input feature depiction of the
II. MACHINE LEARNING TASKS learned function. The accuracy of the
learned function relies sturdily on how the
Machine learning tasks are typically
input object is represented. Normally, the
classified into three broad categories,
input object is altered into a feature vector
depending on the nature of the learning
that contains a number of features that are
"signal" or "feedback" available to a
descriptive of the object. The number of
learning system.
features should not be too large.

1549
All Rights Reserved © 2017 IJARCET
 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
(iv) Decide the structure of the learned
function and corresponding learning
algorithm.
(v) Complete the design. Run the learning
algorithm on the gathered training set. Some
supervised learning algorithms need the user
to find out certain control parameters.
(vi) Assess the accuracy of the learned
function. After parameter adjustment and
learning, the performance of the resulting
function should be measured on a test set
that is separate from the training set.
Unsupervised learning: It is the machine
learning task of inferring a function to depict
concealed structure from "unlabeled" data.
Since the examples specified to the learner
are unlabeled, there is no assessment of the
accuracy of the structure that is output by
the relevant algorithm—which is one way of
distinguishing unsupervised learning from
supervised learning and reinforcement
learning. A central case of unsupervised
learning is the problem of density estimation
in statistics [1].
Reinforcement learning: A computer
program interacts with a vibrant
environment in which it must perform a
certain goal. The program is provided
feedback in terms of rewards
All Rights Reserved © 2017 IJARCET
Volume 6, Issue 10, October 2017, ISSN: 2278 – 1323
punishments as it navigates its problem
space.
III. MACHINE LEARNING
ALGORITHMS
There are number of machine learning
algorithms such as Linear Regression,
Logistic Regression, Decision Tree, SVM
[2], and KNN. Linear Regression is used to
estimate real values (cost of houses, number
of calls, total sales etc.) based on continuous
variable(s). Here, we establish relationship
between independent and dependent
variables by fitting a best line. Logistic
Regression is used to estimate discrete
values based on given set of independent
variable(s). In simple words, it predicts the
probability of occurrence of an event by
fitting data to a logit function. Decision Tree
is a type of supervised learning algorithm
that is mostly used for classification
problems.SVM is a classification method. In
this algorithm, we plot each data item as a
point in n-dimensional space (where n is
number of features you have) with the value
of each feature being the value of a
particular coordinate. K nearest neighbors is
a simple algorithm which stores the entire
available cases and classifies new cases by a
majority vote of its k neighbors.
and
1550
 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 6, Issue 10, October 2017, ISSN: 2278 – 1323

V.CONCLUSION

The article illustrates the concept of machine

learning with its tasks and applications. The
article also highlights the various types of
learning such as supervised learning,
unsupervised learning and reinforcement
learning. In this article a detailed procedure
for solving a problem using supervised
learning has also been discussed..

VI. REFERENCES

1. Talwar, A. and Kumar, Y., 2013. Machine

Learning: An artificial intelligence methodology.
International Journal of Engineering and Computer
Science, 2, pp.3400-3404.

2. Muhammad, I. and Yan, Z., 2015. Supervised

Machine Learning Approaches: A Survey. ICTACT

Fig.1: Machine learning algorithms Journal on Soft Computing, 5(3).

3. Singh, S., Kumar, N. and Kaur, N., 2014. Design

IV. MACHINE LEARNING
Anddevelopment Of Rfid Based Intelligent Security
APPLICATIONS System. International Journal of Advanced Research
in Computer Engineering & Technology (IJARCET)
Machine learning algorithms are widely
Volume, 3.
used in variety of applications like digital 4. Sharma, D., Pabby, G. and Kumar, N., Challenges
image processing(image recognition)[5], big Involved in Big Data Processing & Methods to Solve
data analysis[4], Speech Recognition, Big Data Processing Problems.IJRASET,5(8),pp.841-

Medical Diagnosis, Statistical Arbitrage, 844.

5. Kumar, N. and Gupta, S., 2016. Offline
Learning Associations, Classification,
Handwritten Gurmukhi Character Recognition: A
Prediction etc.
Review. International Journal of Software
Engineering and Its Applications, 10(5), pp.77-86.

1551
All Rights Reserved © 2017 IJARCET
 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 6, Issue 10, October 2017, ISSN: 2278 – 1323

Ms. Diksha completed her

B.Tech from Chitkara University, Himachal
Pradesh in the stream of Electronics and
Communication Engineering. She is now
planning to pursue Masters in science from
abroad.

Mr. Neeraj Kumar is

presently working as Assistant Professor in
Electronics and Communication
Engineering Department at Chitkara
University, Himachal Pradesh, India. He has
more than 6 years of teaching experience.
His area of interest is digital image
processing, digital signal processing.

1552
All Rights Reserved © 2017 IJARCET
View publication stats