Statistics

Sathe, Y. S. (1977) found a new set of sharp upper and lower bounds on the variance of the UMVUE in the case of inverse sampling motivated by the fact that a close-form expression to capture the actual variance is unavailable. Sahai, Ajit (1985) improved the variance upper bound in Sathe, Y. S. (1977). This note is motivated by their papers and by their inherent desire to capture that variance possibly more closely using their set of the sharp Sahai-Sathe’s upper-and-lower bounds. It was very heartening for the authors to realize that the same is possible. In fact the seminal result achieved by the authors in this note could be used for improving the sharpness of these bounds iteratively till it pleases the one using it for the purpose of being close to the actual variance of the UMVUE in the absence of its capture in the closed-form. The achievement is briefly illustrated through a modest empirical study to bring forth the power of the proposed iterative algorithm for the aforesaid purpose.

This study investigates the effects of malaria drugs on children of different gender under the age of five in Nigeria (Umuahia in Abia State as a case study). It was carried out by applying descriptive statistical methods and Poisson log linear regression model on different malaria drugs. To achieve this, five age groups in months (<12, 12-23, 24-35, 36-47, 48-59) and gender (male and female) were used. Secondary datasets were collected from Nigeria Demographic and Health Survey (NDHS) 2013 and department of Health and Human Services, Abia State University, Uturu. The result shows that the rate of usage of all malaria drugs considered in this study is similar for all age groups. However more males than females received Artemisinin-Based Combination Therapy (ACT). The only significant difference found is between the Amodiaquine and Chloroquine, The study therefore concludes that the effect of malaria drugs is dependent on age and gender. Apart from all the main effects of all the variables being significant, there is also a statistically significant interaction between gender, age and drug, p = 0.00134. The poissson regression result predict that in the last 12months, there is a 4.9% and 6.7% increase in the number of gender and drugs respectively considered yearly with (95% CI. 0.281 to 1.678 and 0.006 to 1.033), and also a statistical significant results, P-value = 0.010 & 0.004 respectively.

In this article we express the wigner distribution function in convenient differential forms. In addition We find the temporal evolution of the distribution function for charged particles in an external magnetic field . As a special case we report the wigner function for uniform magnetic field.

The presence of missing values in air quality data is a common issue and important task to deal with it. The data was collected using automated machines which contain missing value that can cause bias result due to differences between observed and predicted data. This paper considers hourly air pollutant concentrations such as NO, NO2, CO, SO2 and meteorological variables such as AT, WS, RH at Kemaman over a four-year period. This paper discusses the performance of Multiple Linear Regression model by using different imputation methods such as mean top bottom, linear interpolation and Markov Chain Monte Carlo (MCMC). The result shows that the MCMC method to replace the missing values obtained the highest R2 (0.649) after fitting to multiple linear regression model for prediction of ozone concentrations with selected precursors. Therefore, MCMC method produced a good fit multiple linear regression for prediction of ozone concentration in this study.

This paper introduces a new improved ratio estimator for population mean in stratified random sampling using calibration estimation theory. Following the results of an empirical study, it is deduced that the proposed estimator is substantially more efficient than existing estimators of its class.

This Article introduces the basic concepts, terminology, goals and procedures underlying the proper statistical design of experiment. This section describes in detail the process of choosing an experimental design. The choice of design in agricultural field experiments depends on the objectives of the experiment and the number of factors to be investigated.

This paper provides a unified framework for chain ratio-product estimation and proposes a generalized chain ratio-product estimator for estimating population mean in stratified sampling. The bias and variance expressions of the proposed estimator have been derived under large sample approximation. Asymptotic optimum estimator (AOE) and its approximate variance estimator are derived with conditions for allowable departure identified. Some existing estimators in theory are obtained and shown to be particular members of the proposed estimator under certain restrictions. Analytical comparisons of the AOE with other existing estimators showed that the AOE is substantially more efficient under certain realistic conditions.

Information criterion, KIC, for the model selection based on kullback-leibler risk symmetric by cavanagh, be present for large sample. KIC as an asymptotically unbiased estimator kullback-leibler risk symmetric, is consider divergency between the true model and the candidate model. It is an inconsistent information criterion. All the criteria that exsist based on kullback-leibler risk (symmetric or asymmetric)are inconsistent. In this article based on information criterion KIC is defined consistent information criterion, MIC. This criterion, MIC also is made based on kullbake-leibler symmetric. At the end these two information criteria for linear regression models has been consider by simulation monte carlo.

Uniformity trials are needed to determine suitable shape and size of the plot for fertility variation in land. The suggested model is more adequate model over well known Fairfield Smith variance law and also removes all the objections by Cochran (1977) for Fairfield Smith variance law model. The adequacy of suggested model has been examined by the given data in Haque et al. (1988). Key words: Fairfield Smith’s Variance Law, nonlinear model, uniformity trial experiments.

This paper was taken the loss and risk concepts, the objective of this paper is to estimate the loss and the risk in the Bayes method, as illustrated in the related case Where the prior distribution of the random variable is discrete and continuous, it applied on crop sales data in the continuous case, while it applied by two examples in the discrete case. Then the estimator of the loss function is the mean posterior distribution of the random variable, so it’s the same with median posterior distribution, hence is also the Bayes estimator with respect to a loss function equal to absolute deviation. The Bayes risk is a real value is the average of the loss function, so the smallest estimator who is the best estimator of the Bayes risk.