Sana S. Buhamra

The Box–Jenkins analysis and neural networks: prediction and time series modelling

Abstract Two approaches, namely the Box–Jenkins (BJ) approach and the artificial neural networks (ANN) approach were combined to model time series data of water consumption in Kuwait. The BJ approach was used to predict unrecorded water consumption data from May 1990 to December 1991 due to the Iraqi invasion of Kuwait in August 1990. A supervised feedforward back-propagation neural network was then designed, trained and tested to model and predict water consumption from January 1980 to December 1999. It is interesting to note that the lagged or delayed variables obtained from the BJ approach and used in neural networks provide a better ANN model than the one obtained either blindly in blackbox mode as has been suggested or from traditional known methods.

Lateral Cephalometric Norms for Adolescent Kuwaitis: Soft Tissue Measurements

Objective: The purpose of this study was to establish lateral cephalometric hard tissue norms for adolescent Kuwaitis and to compare them with published norms. Subjects and Methods: Digital lateral cephalograms were made of 162 Kuwaitis (82 boys and 80 girls of mean age 13.27 ± 0.42 years and 13.21 ± 0.43 years, respectively), with almost ideal occlusion. Anatomic landmarks were identified directly on the digital images. Linear and angular measurements were calculated electronically using the Dolphin version 9 software package. Results: The average subject in the sample had a steeper mandibular plane, a more convex profile with a tendency for reduced chin protrusion, and a more protrusive dentition than the norms of the common analysis systems. In addition, the ranges of the skeletal and dentoalveolar parameters were larger than those reported in the above-mentioned norms. Gender differences were limited to maxillary and mandibular length and lower anterior facial height. Conclusion: The present findings indicate that Kuwaiti norms for incisor inclination and protrusion should be used as a reference when making the extraction decision in Kuwaiti orthodontic patients, and that the variation in skeletal relationships among subjects with satisfactory occlusal compensations is larger than previously documented, suggesting a need for establishing different norms for different skeletal patterns.

Inference concerning quantile for left truncated and right censored data

The problem of both testing and estimating the quantile function when the data are left truncated and right censored (LTRC) is considered. The aim of this communication is two-fold. First, a large sample test statistic to test for the quantile function under the LTRC model is defined and its null and non-null distributions are derived. A Monte Carlo simulation study is conducted to assess the power of the proposed test statistic that is used to define the estimators. Secondly, an improved estimation of the quantile function is investigated. In the spirit of the shrinkage principle in parameter estimation, three estimators assuming an uncertain prior non-sample information on the value of the quantile are proposed. The asymptotic bias and mean square error of the estimators are derived and compared with the usual estimator. The method is illustrated with hypothetical data as well as real data.

Nonparametric inference strategies for the quantile functions under left truncation and right censoring

This paper addresses the problem of estimating the quantile function in a multiple-sample set up when the data are left-truncated and right-censored (LTRC). Assuming an uncertain prior non-sample information on the value of the quantile, we propose improved estimators based on Stein-type shrinkage estimators. A test statistic is also proposed to define improved estimators for the quantile function. The asymptotic bias and risk of the estimators are derived and compared with the benchmark estimator analytically. For several choices of parameters, a Monte Carlo simulation experiment is conducted to appraise the risk reduction of the proposed estimators at different levels of censoring and truncation. We demonstrate that the proposed estimators have superior performance in terms of risk reduction over the benchmark estimator.

Factors for Job Satisfaction among Primary Care Physicians in Kuwait

Objective: To explore factors affecting job satisfaction among primary care physicians in Kuwait amidst the various challenges in the aftermath of the Gulf war. These challenges are in terms of lower number of physicians and lesser number of reopened primary care centres post liberation. Methods: A self-administered questionnaire covering various aspects of job satisfaction was prepared and distributed to a random sample of 185 primary care physician across the five governorates of Kuwait. Factor analysis was employed to ascertain factors underlying job satisfaction. Results: Nine common factors underlying job satisfaction explained 66.9% of the variance in the data. Generally, the physicians were dissatisfied with three factors: salary and promotion, relationship with superiors, and availability of resources. They were satisfied with the remaining six factors: intrinsic work factors, work surroundings, job enthusiasm, relationship with hospital doctors, relationship with colleagues, and finally comfort at work. These nine factors were related to certain demographic variables that were indicative of social structure. Kuwaiti non-family physicians were least dissatisfied with salary and promotion and non-Kuwaiti family physicians were most dissatisfied with this factor. Conclusion: Nationality and specialty proved to be powerful predictors of satisfaction as well as other variables. This is expected because of the advantages Kuwaities have over non-Kuwaities in terms of salary and promotion, and the disadvantages facing family practitioners being, in general, better qualified to do the job, though not appreciated.

Rank tests for two change points

We consider the problem of testing the null hypothesis of no change against two-change points alternatives in a series of independent observations. We propose some Kruskal-Walis-type tests and give their asymptotic null distributions. We also give approximations of their limiting critical values and tables of small sample Monte Carlo critical values. We conducted Monte Carlo simulation studies to compare the powers of the proposed tests with their competitors.

Estimating and testing effect size from an arbitrary population

In this article, we develop inference tools for an effect size parameter in a paired experiment. A class of estimators is defined that includes natural, shrinkage and shrinkage preliminary test estimators. The shrinkage and preliminary test methods incorporate uncertain prior information on the parameter. This information may be available in the form of a realistic guess on the basis of the experimenter’s knowledge and experience, which can be incorporated into the estimation process to increase the efficiency of the estimator. Asymptotic properties of the proposed estimators are investigated both analytically and computationally. A simulation study is also conducted to assess the performance of the estimators for moderate and large samples. For illustration purposes, the method is applied to a data set.

The analysis of VOCs survey data from residences in Kuwait

In this paper, we analyze air samples of volatile organic compounds (VOCs) as well as information on house and source characteristics from 99 residences in Kuwait. The objective of the analysis was two-fold: first to compare between indoor and outdoor concentrations and, secondly, to identify potential sources of pollutants that cause the emission of the VOCs. Since the data are left censored and contain extreme values, traditional methods of analysis are inappropriate and, instead, a nonparametric method due to Akritas (1992, Statistics and Probability Letters, 13, 209–221) was used. Furthermore, the results were compared with those obtained from using the Wilcoxon test and the log rank test.

A combination of Box-Jenkins analysis and neural networks to model and predict water consumption in Kuwait

Two approaches, namely Box-Jenkins approach and artificial neural networks approach (ANN) are combined to model time series data of water consumption in Kuwait. The Box-Jenkins approach was used to predict unrecorded water consumption data from May 1990 to December 1991 due to the Iraqi invasion of Kuwait in August 1990. A supervised feedforward backpropagation neural network was then designed, trained and tested to model and predict water consumption from January 1980 to December 1999. It is interesting to note that the lagged or delayed variables obtained from the Box-Jenkins approach and used in neural networks provide a better ANN model than the one obtained either blindly in blackbox mode as has been suggested or from traditional known methods.

Testing for a change in repeated measures data

Four nonparametric test statistics for the change-point problem with repeated measures data are proposed. In a Monte Carlo simulation study, critical values for the proposed test statistics are simulated and the performances of the proposed tests are compared with the performances of some competitive tests in terms of asymptotic behavior and power. We provide appropriate recommendations for different occurrences of the change-point and illustrate the testing methods using a set of real data.