IRJET - Review on Machine Learning
0 likes18 views
The document discusses machine learning, including its definition, types, applications, and challenges. It provides the following key points: - Machine learning is a field of computer science that allows computers to learn without being explicitly programmed. It is a form of artificial intelligence. - There are three main types of machine learning: supervised learning, unsupervised learning, and reinforcement learning. Supervised learning uses labeled input/output data for training, unsupervised learning finds patterns in unlabeled data, and reinforcement learning learns from interactions. - Common applications of machine learning include virtual personal assistants, social media services, marketing/sales, GPS navigation, weather forecasting, and more. It is used widely on the internet and in industry.
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 3066
REVIEW ON MACHINE LEARNING
Arish Venkat.M.B1
1 Student, B.Sc SS, Sri Krishna Arts and Science College, Tamil Nadu, India.
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract Machine learning is generally a field of computer
science which gives the ability to learn without the use of
programmer. Machine learning is also said to be artificial
intelligence. In this algorithms can be easily understand and
need number of raw data to work according to the set of
algorithms. It can be easily organized and automaticallysolve
more complex data in the problems. It helps in delivering
faster and more accurate results. Some of the programs are
based on internet oriented for example Google maps, amazon
and other online applications. Mainly machinelearning isused
in internet of things. There are three different stages in
machine learning so that it can execute according to that
stages and learns from trainer. There are some of the
challenges in Machine learning which can be solved, but few
things can’t be solved. So machine learningisimportantinday
today’s life. In this paper you will come to know about what is
machine learning, stages, applications andchallengesfaced in
it.
Key Words: machine learning, applications,
challenges,etc..
1. INTRODUCTION
Machine Learning (ML) is a task done based on algorithms
which are used in computer system to perform some of the
instructions. It is a subfield of artificial intelligence.The main
goal of machine learning is to set the structure of data into
the system so that people can utilize system without the use
of coding. Machine learningisusedbycomputerstocalculate
or solve problems. In modern world machine algorithms are
popularly used in private and public sectors. Machine
learning facilities are important in modern computer
systems so that the structure of data we set in the system
will be automated according to the input of data we enter
into computer system. MachineLearning(ML)isregarded as
one of the most promising methodological approaches to
perform network-data analysis and enable automated
network self-configurationandfaultmanagement[1].By this
machine learning algorithm system can learning by itself to
solve the problems. It performs independently on its own
platform. Machine learning is used to teachmachineshowto
handle the data more efficiently [2].
2. MACHINE LEARNING
Machine Learning is used for high level computer program
or algorithm to set a task in the computer system. Machine
Learning is over viewed on bases of theoretical and
mathematical modeling how to set the process to work. This
is an application of artificial intelligence on which it works
automatically without use of user based on task which is set
as an algorithm. Major concept in Machine Learning is that
system can adapt or learn new data without human
interface. Machine Learning knows how to build program in
automatically. This short of Learning is used mainly in
mathematical modeling of data and operations. Once we
inserted the information of data into the system it can’t be
changed. Machine learning (ML) is tolearninformationfrom
the data to process the system.Itisa computational statistics
used for mathematical operations on computer systems.
Machine learning focuses on the development of computer
programs that can access data and use it learn for
themselves. Machine Learning algorithmisusedinvarietyof
applications so that there will be no difficult in performing
tasks.
Fig: 1. Machine Learning
3. TYPES OF MACHINE LEARNING
Machine Learning is used on computer systems based on
algorithms which are set in the task to perform. There are
three types of Machine Learning:](https://image.slidesharecdn.com/irjet-v7i2658-201130062642/85/IRJET-Review-on-Machine-Learning-1-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 3068
4.4 Travelling
Machine learning predicts while travelling with help of GPS
navigation. Today everyone uses GPS for findingthelocation
and the route to go for the marked destination. Now a day’s
people use cab so that it shows the amount when you mark
the location of destination were you want to drop. This is all
done through the help of machine learning technology
connects to GPS connectivity so that it shows the correct
route and traffic in the places while travel in the route.
Common example app for using these kinds ofusesisGoogle
Maps.
Fig: 2. Machine Learning Process
5. CHALLENGES IN MACHINE LEARNING
There are many challenges in machine learningthefirstand
fare most is understanding which processes need to be
automated to solve the problem but not every problem can
be solved, it is difficult to separate the data for set each
problems that needs to be solved. Quality of data is most
important part in machine learning because data is
important in using the machine learningalgorithm,there are
certain number of data’s like dirty data, incomplete data
which cannot be used in machine learning algorithms, so
that it leads to crack of computer. The better solution is to
have a good data so that the process will be done without
any error, this step should be done before you start the
machine learning. Machine learning infrastructure requires
large number of storage capabilities, so that work load can
be done easy or otherwise it should be upgraded. Having
more algorithm lead to trained over a particular data set for
future data reference. The smarter algorithm leads you to
control it difficulty, you can fit more complex model to a
small amount of data set and make sure your data set is
clean without noisy. It can be deal with different stages of
machine learning. Receiving recommendations arecommon
today, it requires reliable or unreliable which is useless for
machine learning data set to be processed. Having bad data
will be leads to wrong results, when the data set is not
understood properly it results wrong. So that only people
are asked to insert quality of data, those unwanted data will
moves to the garbage which is in the system. Even few
machine learning leads to failure. Many people know to
develop machine learning with good data set so that system
can easily understand. As already said that more number of
storage needed otherwise the collection of data set will
slower. There are certain steps for preparing data set
algorithms. Machine learning looks from outside as a small
step of process inside it requires large number of
frameworks. Machine learning need more number time to
get results because it need to gather data and train the
algorithm according the problems that needs to be solved.
6. CONCLUSION
From the above information we can say that machine
learning is important in daily life, which is useful for every
person. Normally machine learning needs a set of data to be
performed. There are many different steps involved in
machine learning they are first data set should be clear and
quality so that there won’t be a lack of error, then analyzing
the problem according to that data set will be matched with
the models, then the trainer will train the algorithm that
needs to be executed. Machine learning is segregated
according to different stages that depend upon the problem
and data set. Machine learning is used in modern world for
various purposes of usages. Normally people use mobile
phones for various reasons in that machine learning is
updated there. Machinelearningisusedinmobile phones for
navigations, booking of cabs and social media. Not only in
these machine learning used it is also used in business fields
over the network, for example amazon is an onlineshopping
application which runs over network from thisthecustomer
buys the product and promotes theproductbycommentand
likes. So in upcoming modernworldmachinelearningwill be
the most used applications.
References
[1]. An Overview on Application of Machine Learning
Techniques in Optical Networks
Francesco Musumeci, Member, IEEE, Cristina Rottondi,
Member, IEEE, Avishek Nag, Member, IEEE, Irene
Macaluso, Darko Zibar, Member, IEEE, Marco Ruffini, Senior
Member, IEEE, and Massimo
Tornatore, Senior Member, IEEE.
[2]. Machine Learning Algorithms: A Review Ayon Dey
Department of CSE, Gautam Buddha University,
Greater Noida, Uttar Pradesh, India.](https://image.slidesharecdn.com/irjet-v7i2658-201130062642/85/IRJET-Review-on-Machine-Learning-3-320.jpg)
IRJET - Review on Machine Learning
- 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 3066 REVIEW ON MACHINE LEARNING Arish Venkat.M.B1 1 Student, B.Sc SS, Sri Krishna Arts and Science College, Tamil Nadu, India. ---------------------------------------------------------------------***---------------------------------------------------------------------- Abstract Machine learning is generally a field of computer science which gives the ability to learn without the use of programmer. Machine learning is also said to be artificial intelligence. In this algorithms can be easily understand and need number of raw data to work according to the set of algorithms. It can be easily organized and automaticallysolve more complex data in the problems. It helps in delivering faster and more accurate results. Some of the programs are based on internet oriented for example Google maps, amazon and other online applications. Mainly machinelearning isused in internet of things. There are three different stages in machine learning so that it can execute according to that stages and learns from trainer. There are some of the challenges in Machine learning which can be solved, but few things can’t be solved. So machine learningisimportantinday today’s life. In this paper you will come to know about what is machine learning, stages, applications andchallengesfaced in it. Key Words: machine learning, applications, challenges,etc.. 1. INTRODUCTION Machine Learning (ML) is a task done based on algorithms which are used in computer system to perform some of the instructions. It is a subfield of artificial intelligence.The main goal of machine learning is to set the structure of data into the system so that people can utilize system without the use of coding. Machine learningisusedbycomputerstocalculate or solve problems. In modern world machine algorithms are popularly used in private and public sectors. Machine learning facilities are important in modern computer systems so that the structure of data we set in the system will be automated according to the input of data we enter into computer system. MachineLearning(ML)isregarded as one of the most promising methodological approaches to perform network-data analysis and enable automated network self-configurationandfaultmanagement[1].By this machine learning algorithm system can learning by itself to solve the problems. It performs independently on its own platform. Machine learning is used to teachmachineshowto handle the data more efficiently [2]. 2. MACHINE LEARNING Machine Learning is used for high level computer program or algorithm to set a task in the computer system. Machine Learning is over viewed on bases of theoretical and mathematical modeling how to set the process to work. This is an application of artificial intelligence on which it works automatically without use of user based on task which is set as an algorithm. Major concept in Machine Learning is that system can adapt or learn new data without human interface. Machine Learning knows how to build program in automatically. This short of Learning is used mainly in mathematical modeling of data and operations. Once we inserted the information of data into the system it can’t be changed. Machine learning (ML) is tolearninformationfrom the data to process the system.Itisa computational statistics used for mathematical operations on computer systems. Machine learning focuses on the development of computer programs that can access data and use it learn for themselves. Machine Learning algorithmisusedinvarietyof applications so that there will be no difficult in performing tasks. Fig: 1. Machine Learning 3. TYPES OF MACHINE LEARNING Machine Learning is used on computer systems based on algorithms which are set in the task to perform. There are three types of Machine Learning:
- 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 3067 3.1 Supervised Learning This is the first type of Machine Learning, which is the simplest and basic to learn the algorithm and implement it. Supervised Learning is the one where we can see the learning can be guided by a teacher. Supervised Learning is based on functions of input and output of data from the trained data. In this Learning input are observed carefully and matched with the perfect trained output of data. The term itself defines that supervisor mean a teacher.Inthiswe teach or train the machine using data to match correctly.For instance there is a set of folder with certain document in it, by using Supervised Learning algorithmwehavetotrainlike document size, date of completion, alphabetical order and etc.., in which we have to train set of data for the perfect outcome based on the training algorithm. It can also classify the data according to output of variable.Thistypeoflearning can also be said that programmer is an teacher who gives input or gives some set of algorithms to get the output by using the help of training data that is users input. 3.2 Unsupervised Learning The Unsupervised Learning works based on the training of machine will get information about the dataset and it will proceed through algorithm which is set default in the computer will act without the guidance. Here the machine works on bases of pattern and differences in the data set without any hell of training of data. It is opposite to supervised learning algorithm,becauseinthislearningthere is no training of data given to machine. Unsupervised learning can’t identify the features of data particularly,butit can identify the pattern and relationships in the clusters of data. Unsupervised Learning algorithms consent to you to act upon extra intricate doling out errands compared to supervised learning. Unsupported knowledge is utterly special from Supervised Learning. This type of learning is used to find unknown pattern in data. It main task is to match the unlabeled data into to the labeled data which it gets from the input of data, which needs a manual intervention. It can identify the feature only for categorization. 3.3 Reinforcement Learning Reinforcement Learning is a typeofMachineLearning,were decision are taken by software agent to maximize at the particular situation. It takes a agent to find the best path in the specific situation. In simple words reinforcement is said to be that output depends on the current state of input. It is purely dependent on label of sequence. For example like chess and self driving cars. Reinforcementlearningisusedin industrial automation as robotics. It can also be used in the environment where there is no analytical solutions are present. It is also used for Machine learning and data processing. This used for long period of time with maximum number of performance. It is not similar to both supervised and unsupervised learning because both have training of data, but this learning does not need it. Its main disadvantage is too much of reinforcement leads to sort the results. 4. MACHINE LEARNING APPLICATIONS Machine Learning is used in computer systems with a set of algorithms which are to be performed by it. There are many applications in machine learning to improve in business decisions, Forecast weather, social media and many other things. There are some of the applications of machine learning used in day to day’s life. 4.1 Virtual Personal Assistant In machine learning application virtual personal assistant plays a major role in this learning. It assists the user to help for future reference or work to be done when the user applies set of instructions to it. Like “Siri” used in apple and “hey Google” in android mobiles is an example for virtual personal assistant. This works based upon the input of the user instruct the machine. Firstwehavetoactivatethem and give input to access them.Machinelearningisimportantpart which collects information and gives output. Few other applications are speech, speech to text conversion and text to speech conversion. 4.2 Social Media Service Social media is used for personal andprivatenewsfeeds and advertisement on social services.Theseareusedfor noticing, viewing and loving in the social media account is based on machine learning. For example Face book is used in social media to communicate with other friends, to look into their accounts and to know the other friends in face book these are operated with help of machine learning algorithms.Face recognition is used to match the picturewhichisuploadedin social media will match with others too. Another example in social media is Google search box is used on bases of complex machine learning algorithm. This is also used in booking cabs, if we book cab our location will be automatically sent to them with use of learning algorithm and the driver will pick up us. 4.3 Marketing and Sales Machine learning predicts while travelling with help of GPS navigation. Today everyone uses GPS for findingthelocation and the route to go for the marked destination. Now a day’s people use cab so that it shows the amount when you mark the location of destination were you want to drop. This is all done through the help of machine learning technology connects to GPS connectivity so that it shows the correct route and traffic in the places while travel in the route. Common example app for using these kinds ofusesisGoogle Maps.
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 3068 4.4 Travelling Machine learning predicts while travelling with help of GPS navigation. Today everyone uses GPS for findingthelocation and the route to go for the marked destination. Now a day’s people use cab so that it shows the amount when you mark the location of destination were you want to drop. This is all done through the help of machine learning technology connects to GPS connectivity so that it shows the correct route and traffic in the places while travel in the route. Common example app for using these kinds ofusesisGoogle Maps. Fig: 2. Machine Learning Process 5. CHALLENGES IN MACHINE LEARNING There are many challenges in machine learningthefirstand fare most is understanding which processes need to be automated to solve the problem but not every problem can be solved, it is difficult to separate the data for set each problems that needs to be solved. Quality of data is most important part in machine learning because data is important in using the machine learningalgorithm,there are certain number of data’s like dirty data, incomplete data which cannot be used in machine learning algorithms, so that it leads to crack of computer. The better solution is to have a good data so that the process will be done without any error, this step should be done before you start the machine learning. Machine learning infrastructure requires large number of storage capabilities, so that work load can be done easy or otherwise it should be upgraded. Having more algorithm lead to trained over a particular data set for future data reference. The smarter algorithm leads you to control it difficulty, you can fit more complex model to a small amount of data set and make sure your data set is clean without noisy. It can be deal with different stages of machine learning. Receiving recommendations arecommon today, it requires reliable or unreliable which is useless for machine learning data set to be processed. Having bad data will be leads to wrong results, when the data set is not understood properly it results wrong. So that only people are asked to insert quality of data, those unwanted data will moves to the garbage which is in the system. Even few machine learning leads to failure. Many people know to develop machine learning with good data set so that system can easily understand. As already said that more number of storage needed otherwise the collection of data set will slower. There are certain steps for preparing data set algorithms. Machine learning looks from outside as a small step of process inside it requires large number of frameworks. Machine learning need more number time to get results because it need to gather data and train the algorithm according the problems that needs to be solved. 6. CONCLUSION From the above information we can say that machine learning is important in daily life, which is useful for every person. Normally machine learning needs a set of data to be performed. There are many different steps involved in machine learning they are first data set should be clear and quality so that there won’t be a lack of error, then analyzing the problem according to that data set will be matched with the models, then the trainer will train the algorithm that needs to be executed. Machine learning is segregated according to different stages that depend upon the problem and data set. Machine learning is used in modern world for various purposes of usages. Normally people use mobile phones for various reasons in that machine learning is updated there. Machinelearningisusedinmobile phones for navigations, booking of cabs and social media. Not only in these machine learning used it is also used in business fields over the network, for example amazon is an onlineshopping application which runs over network from thisthecustomer buys the product and promotes theproductbycommentand likes. So in upcoming modernworldmachinelearningwill be the most used applications. References [1]. An Overview on Application of Machine Learning Techniques in Optical Networks Francesco Musumeci, Member, IEEE, Cristina Rottondi, Member, IEEE, Avishek Nag, Member, IEEE, Irene Macaluso, Darko Zibar, Member, IEEE, Marco Ruffini, Senior Member, IEEE, and Massimo Tornatore, Senior Member, IEEE. [2]. Machine Learning Algorithms: A Review Ayon Dey Department of CSE, Gautam Buddha University, Greater Noida, Uttar Pradesh, India.