Cevikalp H.,Eskiehir Osmangazi University
IEEE Transactions on Pattern Analysis and Machine Intelligence | Year: 2017
In this paper, we propose novel methods that are more suitable than classical large-margin classifiers for open set recognition and object detection tasks. The proposed methods use the best fitting hyperplanes approach, and the main idea is to find the best fitting hyperplanes such that each hyperplane is close to the samples of one of the classes and is as far as possible from the other class samples. To this end, we propose two different classifiers: The first classifier solves a convex quadratic optimization problem, but negative samples can lie on one side of the best fitting hyperplane. The second classifier, however, allows the negative samples to lie on both sides of the fitting hyperplane by using concave-convex procedure. Both methods are extended to the nonlinear case by using the kernel trick. In contrast to the existing hyperplane fitting classifiers in the literature, our proposed methods are suitable for large-scale problems, and they return sparse solutions. The experiments on several databases show that the proposed methods typically outperform other hyperplane fitting classifiers, and they work as good as the SVM classifier in classical recognition tasks. However, the proposed methods significantly outperform SVM in open set recognition and object detection tasks. © 2016 IEEE.
Tamyurek B.,Eskiehir Osmangazi University
IEEE Transactions on Power Electronics | Year: 2013
This paper presents the design of a high-performance sinusoidal pulsewidth modulation (SPWM) controller for three-phase uninterruptible power supply (UPS) systems that are operating under highly nonlinear loads. The classical SPWM method is quite effective in controlling the RMS magnitude of the UPS output voltages. However, it is not good enough in compensating the harmonics and the distortion caused specifically by the nonlinear currents drawn by the rectifier loads. The distortion becomes more severe at high power where the switching frequency has to be reduced due to the efficiency concerns. This study proposes a new design strategy that overcomes the limitations of the classical RMS control. It adds inner loops to the closed-loop control system effectively that enables successful reduction of harmonics and compensation of distortion at the outputs. Simulink is used to analyze, develop, and design the controller using the state-space model of the inverter. The controller is implemented in the TMS320F2808 DSP by Texas Instruments, and the performance is evaluated experimentally using a three-phase 10 kVA transformer isolated UPS under all types of load conditions. In conclusion, the experimental results demonstrate that the controller successfully achieves the steady-state RMS voltage regulation specifications as well as the total harmonic distortion and the dynamic response requirements of major UPS standards. © 1986-2012 IEEE.
Gorenek B.,Eskiehir Osmangazi University
Minerva Medica | Year: 2013
Atrial fibrillation and sustained ventricular arrhythmias are the most common arrhythmias accompanying acute coronary syndromes. Arrhythmias are associated with worse clinical course and increased risk of in-hospital, short-term and long-term mortality in patients with ST-elevation myocardial infarction. This review summarizes the current knowledge on most prevalent arrhythmias in patients with ACS and their management in intensive coronary care unit.
Anik M.,Eskiehir Osmangazi University
Journal of Alloys and Compounds | Year: 2010
Mg2Ni and MgNi alloys were synthesized by the mechanical alloying and their electrochemical hydrogen storage characteristics were investigated. The XRD peaks indicated that MgNi alloy was mainly in the amorphous characteristics while Mg2Ni alloy was composed of mixture of amorphous and nano-crystalline structures. The development of the amorphous structure in MgNi alloy was very fast and the maximum storage capacity was reached at 15 h ball milling. The maximum storage capacity of Mg2Ni alloy was observable after at least 40 h ball milling. Both the initial discharge capacity and capacity retention rate of MgNi alloy were much greater than those of Mg2Ni alloy. This observation was mainly attributed to the relatively reversible charge/discharge behavior of MgNi alloy. The charge transfer resistance of Mg2Ni alloy was greater than that of MgNi alloy especially at the higher depths of discharges. The presence of free electro-catalytically active Ni particles on MgNi alloy surface was thought as the main reason for the facilitated hydrogen charge transfer reactions on this alloy. © 2009 Elsevier B.V. All rights reserved.
Tat M.E.,Eskiehir Osmangazi University
Fuel Processing Technology | Year: 2011
This study presents the energy, exergy and heat release analysis of a John Deere 4045 T 4.5 L, four-stroke, four-cylinder, turbocharged diesel engine. The engine was run with four different types of fuel: yellow grease methyl ester (YGME); soybean oil methyl ester (SME); and soybean oil methyl ester containing either 0.75 or 1.5 w/w % of the cetane improver 2-ethylhexyl nitrate (SME-0.75%EHN and SME-1.5%EHN, respectively). The engine was tested at 1400 1/min under a full load of 352 Nm. For reliability, the fuels were tested three times, and the mean values were compared using different statistical techniques. The objective in this study was to determine the effect of cetane number and ignition delay on the energy and exergy efficiencies of an internal combustion engine and to compare the results for the types of fuel stated earlier. The average thermal efficiency was approximately 40.5%, and the exergetic efficiency was approximately 37.3%. The mean exergetic efficiencies of the fuels were in the order ψSME > ψSME-0.75%EHN > ψSME-1.5%EHN > ψYGME. There were significant differences among the mean values according to Student's t-tests. It is concluded that a lower cetane number, a longer ignition delay period and a higher level of premixed combustion may increase the exergetic efficiency of a diesel engine. © 2010 Elsevier B.V.
Eroglu S.,Eskiehir Osmangazi University
Physica A: Statistical Mechanics and its Applications | Year: 2013
The empirical law uncovered by Menzerath and formulated by Altmann, known as the Menzerath-Altmann law (henceforth the MA law), reveals the statistical distribution behavior of human language in various organizational levels. Building on previous studies relating organizational regularities in a language, we propose that the distribution of distinct (or different) words in a large text can effectively be described by the MA law. The validity of the proposition is demonstrated by examining two text corpora written in different languages not belonging to the same language family (English and Turkish). The results show not only that distinct word distribution behavior can accurately be predicted by the MA law, but that this result appears to be language- independent. This result is important not only for quantitative linguistic studies, but also may have significance for other naturally occurring organizations that display analogous organizational behavior. We also deliberately demonstrate that the MA law is a special case of the probability function of the generalized gamma distribution. © 2013 Elsevier B.V. All rights reserved.
Cevikalp H.,Eskiehir Osmangazi University
Pattern Recognition Letters | Year: 2010
This study presents two new clustering algorithms for partition of data samples for the support vector machine (SVM) based hierarchical classification. A divisive (top-down) approach is considered in which a set of classes is automatically separated into two smaller groups at each node of the hierarchy. The first algorithm splits the data samples based on a variation of the normalized cuts (NCuts) clustering algorithm wherein the weights of adjacency matrix are modified to utilize class membership in the process. The second algorithm also uses the NCuts clustering; however, it considers the involved classes rather than the individual data samples. It uses the minimum distances between the convex hulls of classes as a distance measure for determining the weights of the graph. Splits are determined for both algorithms based on the eigenvector corresponding to the second smallest eigenvalue of a Laplacian matrix, and it is observed that the proposed algorithms generate well-separated and well-balanced clusters. Unlike other clustering methods used for this purpose, the methods in the present study are found to be more suitable when SVMs are used as base classifiers. As demonstrated in the experiments, the proposed clustering algorithms are integrated into the hierarchical SVM classifiers, which results in significantly improved testing times with a negligible decrease in classification accuracies as compared to the traditional multi-class SVMs. © 2010 Elsevier B.V. All rights reserved.
Anik M.,Eskiehir Osmangazi University
International Journal of Hydrogen Energy | Year: 2012
Mg 1.5Ti 0.3Zr 0.1M 0.1Ni (M = Al, B, C, Fe, Pd) type alloys were synthesized by mechanical alloying and their electrochemical hydrogen storage characteristics were investigated. While Al deteriorated the alloy initial discharge capacity, it improved the alloy cyclic stability. The reversibility in the presence of B was very poor. Addition of B increased the alloy initial discharge capacity but reduced the alloy capacity retention rate. The cyclic stability of Fe-including alloy was much better than that of C-including alloy, although both of them have almost the same initial discharge capacities. Both the alloy initial discharge capacity and the alloy cyclic stability were improved significantly in the presence of Pd. The analysis by the electrochemical impedance spectroscopy revealed that Fe increased the corrosion resistance of the alloy without degrading the capacity retention rate. The hydroxide barrier layer in the presence of Al was predicted to be more porous due to the possible high rate selective dissolution of the disseminated Al-oxides. © 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Iphar M.,Eskiehir Osmangazi University
Tunnelling and Underground Space Technology | Year: 2012
Hydraulic impact hammers are mechanical excavators that can be used in tunneling projects economically under geologic conditions suitable for rock breakage by indentation. However, there is relatively less published material in the literature in relation to predicting the performance of that equipment employing rock properties and machine parameters. In tunnel excavation projects, there is often a need for accurate prediction the performance of such machinery. The poor prediction of machine performance can lead to very costly contractual claims. In this study, the application of soft computing methods for data analysis called artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) to predict the net breaking rate of an impact hammer is demonstrated. The prediction capabilities offered by ANN and ANFIS were shown by using field data of obtained from metro tunnel project in Istanbul, Turkey. For this purpose, two prediction models based on ANN and ANFIS were developed and the results obtained from those models were then compared to those of multiple regression-based predictions. Various statistical performance indexes were used to compare the performance of those prediction models. The results suggest that the proposed ANFIS-based prediction model outperforms both ANN model and the classical multiple regression-based prediction model, and thus can be used to produce a more accurate and reliable estimate of impact hammer performance from Schmidt hammer rebound hardness (SHRH) and rock quality designation (RQD) values obtained from the field tests. © 2011 Elsevier Ltd.
Dinleyici E.C.,Eskiehir Osmangazi University
Expert Review of Vaccines | Year: 2010
One of the great success stories in pneumococcal vaccines is the 7-valent conjugated pneumococcal vaccine (PCV-7); and with wider coverage, the broad prevention it offers would be possible with new vaccines. The 7th International Symposium on Pneumococci and Pneumococcal Diseases (ISPPD-7) was held on the 14-18 March 2010 in Tel Aviv, Israel. During this symposium, clinical and epidemiological features in the post-PCV-7 era, current studies about two new approved PCVs (PHiD-CV and PCV-13) and new data about protein-based vaccines were extensively evaluated. Routine surveillance for pneumococcal infections is essential for all countries whether pneumococcal vaccines are available or not, and long-term results of surveillance help us to understand the PCVs potential benefits. Current epidemiological data about serotype 6C have also been presented from different countries and existing PCVs efficacy against serotype 6C should be evaluated. In the post-PCV7 era, we have some new emerging serotypes and we also need new vaccines or interventions focused on protection for children against serotypes prevalent in the developing world or broad protection across all pneumococcal serotypes. Effectiveness of the new PCVs against nasopharyngeal carriage, acute otitis media and invasive infections need to be investigated in the real world; immunogenicity of investigational protein-based vaccines, are still evaluated in different settings. We will have a chance to hear and discuss all the new achievements in the many aspects related to pneumococcal diseases at the next ISPPD-8, which will be held in Iguau Falls in 2012. © 2010 Expert Reviews Ltd.