Journal of Data Science

Abstract: Multiple binary outcomes that measure the presence or absence of medical conditions occur frequently in public health survey research. The multiple possibly correlated binary outcomes may compose of a syndrome or a group of related diseases. It is often of scientific interest to model the interrelationships not only between outcome and risk factors, but also between different outcomes. Applied and practical methods dealing with multiple outcomes from complex designed surveys are lacking. We propose a multivariate approach based on the generalized estimating equation (GEE) methodology to simultaneously conduct survey logistic regressions for each binary outcome in a single analysis. The approach has the following attrac tive features: 1) It enables modeling the complete information from multiple outcomes in a single analysis; 2) it permits to test the correlations between multiple binary outcomes; 3) it allows of discerning the outcome-specific ef fect and the overall risk factor effect; and 4) it provides the measurement of difference of the association between risk factors and multiple outcomes. The proposed method is applied to a study on risk factors for heart attack and stroke in 2009 U.S. nationwide Behavioral Risk Factor Surveillance System (BRFSS) data.

Abstract: We propose two simple, easy-to-implement methods for obtaining simultaneous credible bands in hierarchical models from standard Markov chain Monte Carlo output. The methods generalize Scheff´e’s (1953) approach to this problem, but in a Bayesian context. A small simulation study is followed by an application of the methods to a seasonal model for Ache honey gathering.

Abstract: For the first time, we propose and study the Kumaraswamy generalized half-normal distribution for modeling skewed positive data. The half-normal and generalized half-normal (Cooray and Ananda, 2008) distributions are special cases of the new model. Various of its structural properties are derived, including explicit expressions for the density function, moments, generating and quantile functions, mean deviations and moments of the order statistics. We investigate maximum likelihood estimation of the parameters and derive the expected information matrix. The proposed model is modified to open the possibility that long-term survivors may be presented in the data. Its applicability is illustrated by means of four real data sets.

Abstract: To analyze skewed data, skew normal distribution is proposed by Azzalini (1985). For practical problems of estimating the skewness parame ter of this distribution, Gupta and Gupta (2008) suggested power normal dis tribution as an alternative. We search for another alternative, named tilted normal distribution following the approach of Marshall and Olkin (1997) to add a positive parameter to a general survival function and taking survival function is of normal form. We have found out different properties of this distribution. Maximum likelihood estimate of parameters of this distribu tion have been found out. Comparison of tilted normal distribution with skew normal and power normal distribution have been made.

Abstract: In this paper we propose a new three-parameters lifetime distribu tion with decreasing hazard function, the long-term exponential geometric distribution. The new distribution arises on latent competing risks scenarios, where the lifetime associated with a particular risk is not observable, rather we observe only the minimum lifetime value among all risks, and there is presence of long-term survival. The properties of the proposed distribution are discussed, including its probability density function and explicit algebraic formulas for its survival and hazard functions, order statistics, Bonferroni function and the Lorenz curve. The parameter estimation is based on the usual maximum likelihood approach. We compare the new distribution with its particular case, the long-term exponential distribution, as well as with the long-term Weibull distribution on two real datasets, observing its poten tial and competitiveness in comparison with an usual lifetime distribu

Abstract: The image de-noising is the process to remove the noise from the image naturally corrupted by the noise. The wavelet method is one among the various methods for recovering infinite dimensional objects like curves, densities, images etc. The wavelet techniques are very effective to remove the noise because of its ability to capture the energy of a signal in few energy transform values. The wavelet methods are based on shrinking the wavelet coefficients in the wavelet domain. This paper concentrates on selecting a threshold for wavelet function estimation. A new threshold value is pro posed to shrink the wavelet coefficients obtained by wavelet decomposition of a noisy image by considering that the sub band coefficients have a gener alized Gaussian distribution. The proposed threshold value is based on the power of 2 in the size 2J × 2 J of the data that can be computed efficiently. The experiment has been conducted on various test images to compare with the established threshold parameters. The result shows that the proposed threshold value removes the noise significantly.

Abstract: This article concerns the Bayesian estimation of interest rate mod els based on Euler-Maruyama approximation. Assume the short term inter est rate follows the CIR model, an iterative method of Bayesian estimation is proposed. Markov Chain Monte Carlo simulation based on Gibbs sam pler is used for the posterior estimation of the parameters. The maximum A-posteriori estimation using the genetic algorithm is employed for finding the Bayesian estimates of the parameters. The method and the algorithm are calibrated with the historical data of US Treasury bills.

Abstract: The present paper addresses the propensity to vote with data from the third and fourth rounds of the European Social Survey. The regression of voting propensities on true predictor scores is made possible by estimates of predictor reliabilities (Bechtel, 2010; 2011). This resolves two major problems in binary regression, i.e. errors in variables and imputation errors. These resolutions are attained by a pure randomization theory that incorporates fixed measurement error in design-based regression. This type of weighted regression has long been preferred by statistical agencies and polling organizations for sampling large populations.

Abstract: We apply model-based cluster analysis to data concerning types of democracies, creating an instrument for typologies. Noting several ad vantages of model-based clustering over traditional clustering methods, we fit a normal mixture model for types of democracy in the context of the majoritarian-consensus contrast using Lijphart’s (1999) data on ten variables for 36 democracies. The model for the full period (1945-1996) finds four types of democracies: two types representing a majoritarian-consensus contrast, and two mixed ones lying between the extremes. The four-cluster solution shows that most of the countries have high cluster membership probabilities, and the solution is found to be quite stable with respect to possible measurement error in the variables included in the model. For the recent-period (1971-1996) data, most countries remain in the same clusters as for the full-period data.

Abstract: When comparing the performance of health care providers, it is important that the effect of such factors that have an unwanted effect on the performance indicator (eg. mortality) is ruled out. In register based studies randomization is out of question. We develop a risk adjustment model for hip fracture mortality in Finland by using logistic regression. The model is used to study the impact of the length of the register follow-up period on adjusting the performance indicator for a set of comorbidities. The comorbidities are congestive heart failure, cancer and diabetes. We also introduce an implementation of the minimum description length (MDL) principle for model selection in logistic regression. This is done by using the normalized maximum likelihood (NML) technique. The computational burden becomes too heavy to apply the usual NML criterion and therefore a technique based on the idea of sequentially normalized maximum likelihood (sNML) is introduced. The sNML criterion can be evaluated efficiently also for large models with large amounts of data. The results given by sNML are then compared to the corresponding results given by the traditional AIC and BIC model selection criteria. All three comorbidities have clearly an effect on hip fracture mortality. The results indicate that for congestive heart failure all available medical history should be used, while for cancer it is enough to use only records from half a year before the fracture. For diabetes the choice of time period is not as clear, but using records from three years before the fracture seems to be a reasonable choice.