Browsing by Subject "Cross validation"

Now showing 1 - 3 of 3

Open Access
Bandwidth selection for kernel density estimation using fourier domain constraints
(Institution of Engineering and Technology, 2016) Suhre, A.; Arıkan, Orhan; Çetin, A. Enis
Kernel density estimation (KDE) is widely-used for non-parametric estimation of an underlying density from data. The performance of KDE is mainly dependent on the bandwidth parameter of the kernel. This study presents an alternative method of estimating the bandwidth by incorporating sparsity priors in the Fourier transform domain. By using cross-validation (CV) together with an l1 constraint, the proposed method significantly reduces the under-smoothing effect of traditional CV methods. A solution for all free parameters in the minimisation is proposed, such that the algorithm does not need any additional parameter tuning. Simulation results indicate that the new approach is able to outperform classical and more recent approaches over a set of distributions of interest.
Open Access
Energy efficient dynamic virtual machine allocation with cpu usage prediction in cloud datacenters
(2018-01) Urul, Gökalp
With tremendous increase in Internet capacity and services, the demand for cloud computing has also grown enormously. This enormous demand for cloud based data storage and processing forces cloud providers to optimize their platforms and facilities. Reducing energy consumption while maintaining service level agreements (SLAs) is one of the most important issues in this optimization effort. Dynamic virtual machine allocation and migration is one of the techniques to achieve this goal. This technique requires constant measurement and prediction of usage of machine resources to trigger migrations at right times. In this thesis, we present a dynamic virtual machine allocation and migration method utilizing CPU usage prediction to improve energy efficiency while maintaining agreed quality of service levels in cloud datacenters. Our proposed method, called LRAPS, tries to estimate short-term CPU utilization of hosts based on their utilization history. This estimation is then used to detect overloaded and underloaded hosts as part of live migration process. If a host is overloaded, some of the VMs running on that host are migrated to other hosts to avoid SLA violations; if a host is underloaded, all of the VMs in that host are tried to be migrated to other machines so that the host can be powered off. We did extensive simulation experiments using CloudSim to evaluate the efficiency and effectiveness of our proposed method. Our simulation experiments show that our method is feasible to apply and can signi cantly reduce power consumption and SLA violations in cloud systems.
Open Access
Recognizing daily and sports activities in two open source machine learning environments using body-worn sensor units
(Oxford University Press, 2014-11) Barshan, B.; Yüksek, M. C.
This study provides a comparative assessment on the different techniques of classifying human activities performed while wearing inertial and magnetic sensor units on the chest, arms and legs. The gyroscope, accelerometer and the magnetometer in each unit are tri-axial. Naive Bayesian classifier, artificial neural networks (ANNs), dissimilarity-based classifier, three types of decision trees, Gaussian mixture models (GMMs) and support vector machines (SVMs) are considered. A feature set extracted from the raw sensor data using principal component analysis is used for classification. Three different cross-validation techniques are employed to validate the classifiers. A performance comparison of the classifiers is provided in terms of their correct differentiation rates, confusion matrices and computational cost. The highest correct differentiation rates are achieved with ANNs (99.2%), SVMs (99.2%) and a GMM (99.1%). GMMs may be preferable because of their lower computational requirements. Regarding the position of sensor units on the body, those worn on the legs are the most informative. Comparing the different sensor modalities indicates that if only a single sensor type is used, the highest classification rates are achieved with magnetometers, followed by accelerometers and gyroscopes. The study also provides a comparison between two commonly used open source machine learning environments (WEKA and PRTools) in terms of their functionality, manageability, classifier performance and execution times. © 2013 © The British Computer Society 2013. All rights reserved.