• Communications for Statistical Applications and Methods
• /
• v.20 no.6
• /
• pp.491-498
• /
• 2013

We consider support vector machines(SVM) to predict Y with p numerical variables $X_1$, ${\ldots}$, $X_p$. This paper aims to build a biplot of p explanatory variables, in which the first dimension indicates the direction of SVM classification and/or regression fits. We use the geometric scheme of kernel principal component analysis adapted to map n observations on the two-dimensional projection plane of which one axis is determined by a SVM model a priori.

• Communications for Statistical Applications and Methods
• /
• v.20 no.1
• /
• pp.1-13
• /
• 2013

This paper introduces an extended version of ${\kappa}$-records. Kullback-Leibler (K-L) information between two generalized distributions arising from ${\kappa}$-records is derived; subsequently, it is shown that K-L information does not depend on the baseline distribution. The behavior of K-L information for order statistics and ${\kappa}$-records, is studied. The exact expressions for K-L information between distributions of order statistics and upper (lower) ${\kappa}$-records are obtained and some special cases are provided.

• Communications for Statistical Applications and Methods
• /
• v.20 no.6
• /
• pp.467-474
• /
• 2013

It is common to measure multiple coverage metrics during software testing. Software reliability growth models and coverage growth functions have been applied to each coverage metric to evaluate software reliability; however, analysis results for the individual coverage metrics may conflict with each other. This paper proposes the virtual coverage metric of a normalized first principal component in order to avoid conflicting cases. The use of the virtual coverage metric causes a negligible loss of information.

• Communications for Statistical Applications and Methods
• /
• v.20 no.6
• /
• pp.427-438
• /
• 2013

Based on progressive Type II censored sampling which is an important method to obtain failure data in a lifetime study, we suggest a very general form of Bayesian prediction bounds from two parameters exponentiated Weibull distribution using the proper general prior density. For this, Markov chain Monte Carlo approach is considered and we also provide a simulation study.

• Communications for Statistical Applications and Methods
• /
• v.20 no.5
• /
• pp.357-366
• /
• 2013

This paper defines an improvement for estimating the population mean of a study variable using auxiliary information and known values of certain population parameter(s), when there is a non-response in a study as well as on auxiliary variables. Under a simple random sampling without a replacement (SRSWOR) scheme, the mean square error (MSE) of all proposed estimators are obtained and compared with each other. Numerical illustration is also given.

• Communications for Statistical Applications and Methods
• /
• v.20 no.3
• /
• pp.169-174
• /
• 2013

The ROC curve is drawn with two conditional cumulative distribution functions (or survival functions) of the univariate random variable. In this work, we consider joint cumulative distribution functions of k random variables, and suggest a ROC curve for multivariate random variables. With regard to the values on the line, which passes through two mean vectors of dichotomous states, a joint cumulative distribution function can be regarded as a function of the univariate variable. After this function is modified to satisfy the properties of the cumulative distribution function, a ROC curve might be derived; moreover, some illustrative examples are demonstrated.

• Communications for Statistical Applications and Methods
• /
• v.20 no.5
• /
• pp.377-385
• /
• 2013

We consider a compound Poisson risk model in which the premium rate changes when the surplus exceeds a threshold. The explicit form of the ruin probability for the risk model is obtained by deriving and using the overflow probability of the workload process in the corresponding M/G/1 queueing model.

• Communications for Statistical Applications and Methods
• /
• v.20 no.3
• /
• pp.225-233
• /
• 2013

The linear support vector machine(SVM) is motivated by the maximal margin separating hyperplane and is a popular tool for binary classification tasks. Many studies exist on the consistency properties of SVM; however, it is unknown whether the linear SVM is consistent for estimating the optimal classification boundary even in the simple case of two Gaussian classes with a common covariance, where the optimal classification boundary is linear. In this paper we show that the linear SVM can be inconsistent in the univariate Gaussian classification problem with a common variance, even when the best tuning parameter is used.

• Communications for Statistical Applications and Methods
• /
• v.20 no.5
• /
• pp.417-425
• /
• 2013

We consider an extended version of a logarithmic series distribution and discuss the estimation of its parameters by the method of moments and the method of maximum likelihood. Test procedures are suggested to test the significance of the additional parameter of this distribution and all procedures are illustrated with the help of real life data sets. In addition, a simulation study is conducted to assess the performance of the estimators.

• Communications for Statistical Applications and Methods
• /
• v.20 no.3
• /
• pp.199-205
• /
• 2013

We employ sieve bootstraps for empirical likelihood tests in time series models because their null distributions are often vulnerable to the presence of serial dependence. We found a significant size refinement of the bootstrapped versions of a Lagrangian Multiplier type test statistic regardless of the bandwidth choice required by long-run variance estimations.