2 Machine Learning Based Presaging Technique For
2 Machine Learning Based Presaging Technique For
Mohan R.M
Associate Professor , Department of CSE
Panimalar Engineering College
Chennai, India
[email protected] m
Abstract— In recent years, Cloud computing is emerging as these third-party service providers. Later the cloud technology
an indispensable technology in handling the increasing number has taken dominance over the market of service providing and
of users and data generated by them. Despite the hype, the the third-party service providers diminished slowly in the
increasing data rate proportionally increases the workload of the following years after the introduction of cloud. The main
cloud service providers in order to meet the consumer demands. reason cloud dominated the world of service providing is its
Moreover this situation results in developing many cloud service user-friendly set of terms which are reliable and always
providers in order deliver the required and efficient services to updated as per the customers’ demands. The cloud technology
the users. But the way to select the appropriate service provider is a vast concept to explore and also complex for the provider
is still remaining as a bottleneck. To solve the increasing server
negotiation challenges, a pattern-based service negotiation
in providing service. There are many technological updates
method has been utilized in this research work. This method has
made and also many systems are proposed to make cloud
been implemented with an improved suggestion feature, which service providers work easier but there are still few
provides users with similar patterns recognized form their past unexplored concepts remained where novel ideas are required
entries to assist the users with their current selection. By to clear the toils faced by the service providers. The primary
incorporating implicit tracking approach, a sequence of the step for our proposed idea starts with the designing of a
services accessed by the users have been maintained and this can system that provides customer services as per their needs or
be leveraged as suggestion to the users. Based on the pattern, the request. The basic customer requests are processed through the
most important services are ranked according to the users’ cloud controller which holds the cloud service provider. The
activity and the data has been observed using machine learning cloud service providers provide a large number of services and
algorithm. The ranked list is given to the user to select the resources through the cloud environment. There are diverged
required service. The proposed technique increases the layers in the service providing like software as a service,
performance of the cloud environment and also the strain which is provided at the presentation or application layer.
handled by the service provider in suggesting the utility pattern Similar to the before services it also manages platform-
for the user is reduced by using the machine learning concept. oriented services which provide services at the platform level
and infrastructural services. These service selections are
Keywords— Cloud services, cloud computing, pattern-based performed through service level agreements. The selection
recognition, machine learning, utility pattern, K nearest neighbor
process is feasible for the user but when it comes to
algorithm.
negotiation of the user selection change or utility it is tough to
be done in a loosely coupled environment such as the cloud.
I. INT RODUCT ION Thus the initial selection performed by the user must be the
The world before the utilization of cloud services was way best one to which we utilize the proposed system. The
different. The companies established were forced to maintain proposed system explains the utilization of a machine
their servers. The companies which were not capable of learning algorithm which tends to analyze the developed
maintaining the servers were dependent on third -party multi-user utility pattern and then we provide the user with a
providers. The service providers had a fragile set of terms and definite set of choices from the study of the pattern of multi-
conditions. This always made companies less dependent on users. Thus we try to get an upgraded cloud service system
Authorized licensed use limited to: University College London. Downloaded on May 25,2020 at 22:51:24 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the Second International Conference on Innovative Mechanisms for Industry Applications (ICIMIA 2020)
IEEE Xplore Part Number: CFP20K58-ART; ISBN: 978-1-7281-4167-1
that helps the user in selecting a better choice from a set of reliability, responsiveness, security, and elasticity. These
functionalities provided by the cloud service provider. The quality dimensions define what the user precisely needs and
proposed system has the potential to incorporate itself in the the service provider can easily provide it to the user.
present-day cloud service providers system to benefit both the
user and provider in various ways. The major advantage in this Paper 3:
system is the better choice selection for the user and also less Source: COMMUN. ACM, vol. 40, Number 4, pp. 50–58. [5]
chance of the possibility of negotiation in the future to the Published in: April, 2010
service provider from the user. Title: A view of cloud computing
Description: The paper defines the functionalities and
II. RELAT ED W ORKS potential the cloud technology can provide for the IT industry
The proposed work have considered a few of the related and as well as software development companies. The paper
works and made them as our base papers to state the proposed also explains about the capability and utility of the cloud
system. Such papers are mentioned below: concept in the present-day world.
Paper 4:
Paper 1:
Source: IEEE Transactions on Cloud Computing, Vol. 3,
Source: Int. J. Intelligent Enterprise, Vol. 4, Nos. ½ [2]
pp.66–79, IEEE. [9]
Published in: 2017
Published in: January–March, 2015
Title: Multi-user utility pattern based cloud service negotiation
Title: ‘SelCSP: a framework to facilitate selection of cloud
for quality of service improvement: An implicit interest
service providers’
prediction approach.
Description: The paper proposes an efficient system to help
Description: As per this paper, it states a solution for
the user to choose a better service provider but does not
developing a better selection system for the user and reduce
mention any sort of precise algorithm or method to eliminate
the negotiation problems but couldn’t completely eradicate
the toils faced by the user to select the cloud service provider
once the toil between the user and service provider starts.
with better options of services.
Paper 2: A. Advantages
Source: IEEE Transactions on Industrial Informatics, Vol. 10,
Number 2, IEEE. [4] Reduced chances of later negotiations from user
Published in: May, 2014 Efficient prediction system from log
Title: CLOUDQUA L: a quality model for cloud services Better and precise results
Description: It mentions about a system called
‘CLOUDQUAL’, which is a six metrics holding service
assistant. The metrics it considers are usability, availability,
Authorized licensed use limited to: University College London. Downloaded on May 25,2020 at 22:51:24 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the Second International Conference on Innovative Mechanisms for Industry Applications (ICIMIA 2020)
IEEE Xplore Part Number: CFP20K58-ART; ISBN: 978-1-7281-4167-1
B. Methodologies Used The best example we can say for supervised learning is a
The concepts we propose utilizing an already existing machine making predictions. The prediction requires pre-
system technology and in addition to it a few methods of labelled data through which it learns and starts to predict from
machine learning. The major theories we adopt from machine the next instance when the new data is given as input and the
learning are the algorithms that aid in solving classification and prediction done by the system is given as output to the given
regression problems. The algorithm we use here is the k nearest input.
algorithm as it is easy to understand and as well as easy to (ii) Regression and classification: In regression type of
implement. Along with the algorithms we also utilized a small problems, we try to predict some continuous -valued output and
part of an existing system. The reason we require the past whereas in classification we predict some discrete values.
system is that the stated system is never a novel methodology
but it’s an updated proposal for the existing system. Thus the (iii) KNN algorithm: This algorithm comes under
existing system available and also suitable for us is a multi-user supervised learning which involves both regression and
utility pattern. The utility pattern is analyzed by the aid of the classification based solutions. The ideology of KNN is that
algorithm to benefit client from choosing a good and as well a similar data points are close to each other. KNN believes in
suitable cloud service provider. such an ideology for the algorithm to be useful.
C. Equations and Processes The above image is an example of the KNN algorithm; we
(i) Supervised machine learning: The basic standard of can observe the above image that the similar data points are
supervised machine learning is that the learning algorithm is close to each other. KNN captures the idea of similarity which
provided with some sample data or a training set from which in some cases can be distance, proximity, or closeness using
the algorithm should learn from. In supervised learning, there some mathematics. (Calculating the distance between points on
are two types of problems defined. The first one is regression graphs)
and another one is classification. Before understanding these
Fundamental KNN Algorithm Steps:
concepts we shall design basic steps of a supervised learning
process which is as follows. 1. Initialize with loading data
2. Declare k with chosen number of neighbors
Figure 2.1 Image displaying similar data points typically
3. For every sample of data
existing close
3.1 Calculate the distance between the query example and
the current example from the data
3.2 Add the distance and index of the example to avail an
ordered collection
4. Sort the ordered collections of distances and indices in
ascending order with respect to distances.
5. Select the first K entries from sorted set
6. Avail the labels of the selected K entries
7. If regression, return the mean of the K labels
8. If classification, return the mode of the K labels
IV. CONCLUSION
After the text edit has been completed, the paper is ready
for the template. The stated system is a proposal to initiate a
novel concept in the existing system to make it provide better
results.. The proposed system is constantly being worked upon
and various other methods and designs are constantly being
Basic Steps: improvised in terms of machine learning concept and also the
updates in the existing system used. Apart from this the
1. Provide training set data into the learning algorithm concept stated has great usage and cloud service users can be
benefitted from it. These pros can be eventually increased by
2. The learning algorithm understands the data and learns
undergoing constant research. Henceforth, the concept of
the hypothesis ‘k’
designing a machine for assessing the cloud service providers
3. Feed a new input data to k ranking through machine learning is observed as a novel
approach and has a great future scope in developing this system
4. This time the algorithm provides output which we can to the next level.
avail from k
Authorized licensed use limited to: University College London. Downloaded on May 25,2020 at 22:51:24 UTC from IEEE Xplore. Restrictions apply.
Proceedings of the Second International Conference on Innovative Mechanisms for Industry Applications (ICIMIA 2020)
IEEE Xplore Part Number: CFP20K58-ART; ISBN: 978-1-7281-4167-1
References
[1] “ A framework for ranking of cloud comput ing services” by Garg S.K.,
Ve rsteeg S. and Buyya R, Future Gener. Comput. Syst., Vol. 29, Number 4,
pp.1012–1023.
[2] “ Multi-user utility pattern based cloud service negotiation for quality of
service improvement: an implicit interest prediction approach Learning” by
Mohan Manoharan and Selvarajan Saraswathi, IJIE, Vol.4 and issue 1-2,
2014.
[3] “ Applications integration in a hybrid cloud computing environment:
modelling and platform” by Li Q. et al., Enterp. Inf. Syst., Vol. 7, Number 3,
pp.237–271.
[4] “ CLOUDQUAL: a quality model for cloud services” by Zheng, X.,
Martin P., Brohman K and Xu L.D, IEEE T ransactions on Industrial
Informatics, May, Vol. 10, Number 2, IEEE.
[5] “ A view of cloud computing” by M. Armbrust et al., Commun. ACM,
vol. 40, Number 4, pp. 50–58, 2010.
[6] “ Enterprise cloud service architectures” by H. Wang, W. He, and F. K.
Wang, Inf. T echnol. Manag., vol. 13, Number 4, pp. 445–454, 2012.
[7] “ Goods, products and services” by G. Parry, L. Newnes, and X. Huang,
Service Design and Delivery, Service Science: Research and Innovations in
the Service Economy, M. Macintyre et al., Eds. New York, NY, USA:
Springer, 2011, pp. 19–29.
[8] “ T owards a trust management system for cloud computing” by S. K.
Habib, S. Ries, M. Muhlhauser, Proc. IEEE 10th Int. Conf. T rust Secur.
Privacy Comput. Commun., pp. 933-939, 2011.
[9] “ SelCSP: a framework to facilitate selection of cloud service providers”
by Ghosh N., Ghosh S.K. and Das S.K., IEEE T ransactions on Cloud
Computing, January–March, Vol. 3, Number 1, pp.66–79, IEEE.
[10] “ Qos-aware data replication for data intensive applications in cloud
computing systems”by J. Lin, C. Chen, J. Chang, IEEE T rans. Cloud
Comput., vol. 1, Number 1, pp. 101-115, Jan.–Jun. 2013.
[11] “ A trusted computing environment model in cloud architecture” by X. Li,
L. Zhou, Y. Shi, Y. Guo, Proc. Int. Conf. Mach. Learn. Cybern., vol. 6, pp.
2843-2848.
[12] “ Cloud provider transparency: An empirical evaluation” by W. A.
Paule y, IEEE Security Privacy, vol. 8, Number 6, pp. 32-39, 2010.
[13] Raj, J. S., & Ananthi, J. V. (2019) “RECURRENT NEURAL
NET WORKS AND NONLINEAR PREDICT ION IN SUPPORT VECT OR
MACHINES”, Journal of Soft Computing Paradigm (JSCP), 1(01), 33 -40.
Authorized licensed use limited to: University College London. Downloaded on May 25,2020 at 22:51:24 UTC from IEEE Xplore. Restrictions apply.