Yildiz O.T.,Isik University
IEEE Transactions on Knowledge and Data Engineering | Year: 2013
Rule learning algorithms, for example, RIPPER, induces univariate rules, that is, a propositional condition in a rule uses only one feature. In this paper, we propose an omnivariate induction of rules where under each condition, both a univariate and a multivariate condition are trained, and the best is chosen according to a novel statistical test. This paper has three main contributions: First, we propose a novel statistical test, the combined 5 × 2 cv t test, to compare two classifiers, which is a variant of the 5 × 2 cv t test and give the connections to other tests as 5 × 2 cv F test and k-fold paired t test. Second, we propose a multivariate version of RIPPER, where support vector machine with linear kernel is used to find multivariate linear conditions. Third, we propose an omnivariate version of RIPPER, where the model selection is done via the combined 5 × 2 cv t test. Our results indicate that 1) the combined 5 × 2 cv t test has higher power (lower type II error), lower type I error, and higher replicability compared to the 5 × 2 cv t test, 2) omnivariate rules are better in that they choose whichever condition is more accurate, selecting the right model automatically and separately for each condition in a rule. © 1989-2012 IEEE.
Ozaydin F.,Isik University
International Journal of Theoretical Physics | Year: 2015
Recent studies on quantum Fisher information (QFI) have been focused mostly on qubit systems within the context of how entanglement helps surpassing the classical limit of separable states and the limit that a given entangled system can achieve for parameter estimation. However, there are only a few works on bound entangled systems. In this work, we study the QFI of a system of the smallest dimension that bound entanglement can be observed: A bipartite quantum system of two particles of three-levels each. An interesting property of this state is that depending only on a parameter, the state can be separable, bound entangled or free entangled. We show that QFI exhibits a smooth and continues increase with respect to this parameter throughout the transition from separable to bound entangled and from bound entangled to free entangled regions. We show that in any region, this state is not useful for sub-shot noise interferometry. We also relate the QFI of this state with its geometric discord and show how these two properties exhibit a similar behavior throughout this transition. © 2015, Springer Science+Business Media New York.
Aksezer C.S.,Isik University
Engineering Failure Analysis | Year: 2011
Reliability is an important aspect of product perception and manufacturers are compelled to take corrective actions on the items failing within the warranty period. Automotive manufacturers are being exposed to significant operating costs as a result of warranty claims affecting an individual unit or mandatory (sometimes voluntary) recalls affecting a batch. Underlying principles of warranty modeling are built by considering both subjective issues and objective constraints such as competition, quality, and performance under the goal of achieving desired levels of reliability and cost in a balanced manner. This paper reviews the warranty cost models with an emphasis on the failure analysis of used vehicles. Expected warranty costs are calculated by taking into account the age, usage, and maintenance data of the product in question. Failure intensities and characteristics are identified in order to propose a policy that highlights the trade-off between the cost and the warranty length. A case study on a popular brand's initiation of factory certified pre-owned program for the local automobile market of Turkey is presented in detail. © 2011 Elsevier Ltd.
Bayindir C.,Isik University
Physical Review E - Statistical, Nonlinear, and Soft Matter Physics | Year: 2016
In this paper we study the properties of the chaotic wave fields generated in the frame of the Kundu-Eckhaus equation (KEE). Modulation instability results in a chaotic wave field which exhibits small-scale filaments with a free propagation constant, k. The average velocity of the filaments is approximately given by the average group velocity calculated from the dispersion relation for the plane-wave solution; however, direction of propagation is controlled by the β parameter, the constant in front of the Raman-effect term. We have also calculated the probabilities of the rogue wave occurrence for various values of propagation constant k and showed that the probability of rogue wave occurrence depends on k. Additionally, we have showed that the probability of rogue wave occurrence significantly depends on the quintic and the Raman-effect nonlinear terms of the KEE. Statistical comparisons between the KEE and the cubic nonlinear Schrödinger equation have also been presented. © 2016 American Physical Society.
Arik S.,Isik University
IEEE Transactions on Neural Networks and Learning Systems | Year: 2014
In this paper, we study the global asymptotic robust stability of delayed neural networks with norm-bounded uncertainties. By employing the Lyapunov stability theory and homeomorphic mapping theorem, we derive some new types of sufficient conditions ensuring the existence, uniqueness, and global asymptotic stability of the equilibrium point for the class of neural networks with discrete time delays under parameter uncertainties and with respect to continuous and slope-bounded activation functions. An important aspect of our results is their low computational complexity, as the reported results can be verified by checking some properties of symmetric matrices associated with the uncertainty sets of the network parameters. The obtained results are shown to be generalizations of some of the previously published corresponding results. Some comparative numerical examples are also constructed to compare our results with some closely related existing literature results. © 2012 IEEE.