Classification Techniques in Machine Learning: Applications and Issues

Journal of Basic & Applied Sciences, Journal Year: 2017, Volume and Issue: 13, P. 459 - 465

Published: Jan. 5, 2017

Classification is a data mining (machine learning) technique used to predict group membership for instances. There are several classification techniques that can be purpose. In this paper, we present the basic techniques. Later discuss some major types of method including Bayesian networks, decision tree induction, k-nearest neighbor classifier and Support Vector Machines (SVM) with their strengths, weaknesses, potential applications issues available solution. The goal study provide comprehensive review different in machine learning. This work will helpful both academia new comers field learning further strengthen basis methods.

Language: Английский

---

Structure learning in Bayesian Networks using regular vines

Computational Statistics & Data Analysis, Journal Year: 2016, Volume and Issue: 101, P. 186 - 208

Published: March 10, 2016

Language: Английский

---

Machine learning based Call Admission Control approaches: A comparative study

Published: Oct. 1, 2010

The importance of providing guaranteed Quality Service (QoS) cannot be overemphasised, especially in the NGN environment which supports converged services on a common IP transport network. Call Admission Control (CAC) mechanisms do provide QoS to class-based proactive manner. However, due factors complexity, scale and dynamicity NGN, Machine Learning techniques are favoured analytical approaches for autonomous CAC. This paper is an effort compare performance two such - Neural Networks (NN) Bayesian (BN), model network behaviour estimate metrics used CAC algorithm. It provides way find optimum training size accurate predictions. Performance comparison based wide range experiments through simulated Opnet. outcome this comparative study some interesting insights into NN BN models how they can utilised better implementations.

Language: Английский

---

An Autonomic Open Marketplace for Inter-Cloud Service Management

Published: Dec. 1, 2011

The rise of utility in cloud computing and telecommunications has introduced greater complexity the provisioning performance management remote services. We propose extended strategies for this complexity. Our overall aim is to accept responsibility complex service an open marketplace. Responsibility is, firstly, defined by aiming cover totality modern services, managing both connectivity virtual infrastructure. Secondly, further as risk resilience operation service. With these aims, we are working towards a bundled provider agent architecture, which can negotiate on market. This approach aims also optimise utilisation providers infrastructure while reducing failure users through total management. present specification, design simulation agents marketplace environment.

Language: Английский

---

Ismael: Using Machine Learning to Predict Acceptance of Virtual Clusters in Data Centers

IEEE Transactions on Network and Service Management, Journal Year: 2019, Volume and Issue: 16(3), P. 950 - 964

Published: July 12, 2019

Existing virtual network admission control algorithms targeting high utilization of data center infrastructure are computationally expensive or provide poor performance. In particular, existing have in common that they oblivious to the past, i.e., requests handled a fire-and-forget manner, not taking into account information from previously solved instances. This can be inefficient and misses out on basic optimization opportunity: as for any algorithm faces repeating problem instances, it may beneficial learn states outcome acceptance decisions past. this paper, we propose Ismael, machine learning framework predicting clusters, one most abstractions centers. Ismael configured with, from, different by combining fixed-size feature representations graphs with convolutional neural fully connected deep network. We report extensive simulations, which demonstrate is possible mimic existing, intensive an accuracy up 94 %, while significantly reducing runtime.

Language: Английский

---

A Bayesian Approach to Service Selection for Secondary Users in Cognitive Radio Networks

International Journal of Advanced Computer Science and Applications, Journal Year: 2015, Volume and Issue: 6(10)

Published: Jan. 1, 2015

In cognitive radio networks where secondary users (SUs) use the time-frequency gaps of primary users' (PUs) licensed spectrum opportunistically, experienced throughput SUs depend not only on traffic load PUs but also PUs' service type. Each has its own pattern channel usage, and if know dominant then they can make a better decision choosing which is to be used at specific time get best advantage channel, in terms higher achievable throughput. However, it difficult inform directly services each area, for practical reasons. This paper proposes learning mechanism embedded sense length time. algorithm recommends upon sensing free choose order performance, maximum achieved minimum delay. The proposed based Bayesian approach that predict performance requested given SU. Simulation results show this selection method outperforms blind opportunistic SU selection, significantly.

Language: Английский

---

An autonomic open marketplace for service management and resilience

Conference on Network and Service Management, Journal Year: 2011, Volume and Issue: unknown, P. 417 - 421

Published: Oct. 24, 2011

Expansion in telecommunications services, such as triple play and unified communications, introduces complexity that adversely affects service network provisioning, especially terms of provisioning times the risk delivery (failure) new services. We envision a marketplace which all manner complex services will be provisioned, their performance managed, against poor performance. The first phase our work is focus on architecture, negotiation management, lead to effective specification management requirements. are working towards bundled agent can negotiate an open single market, eventually help optimise utilisation providers networks while reducing failure users. Our date has been specification, behaviour, definition simulation agents for delivery.

Language: Английский

---

CARMA: Complete autonomous responsible management agents for telecommunications and inter-cloud services

Published: April 1, 2012

The continuing rise in telecommunication and cloud services usage is matched by an increased complexity maintaining adequate performance management. To combat this complexity, researchers companies are exploring a variety of management strategies to leverage their individual infrastructures provide better utilisation. We extend these addressing the complexities that arise through interaction multiple providers when providing modern complex service. Our overall aim for accept responsibility service open marketplace. Responsibility is, firstly, defined aiming cover totality services, managing both connectivity virtual infrastructure. Secondly, as risk resilience provisioning operation With aims, we working towards bundled provider agent architecture, which can negotiate on market. This approach aims also optimise utilisation infrastructure while reducing failure users total present specification, design simulation Complete Autonomous Responsible Management Agents (CARMA) marketplace environment.

Language: Английский

---