The Dhaka University Journal of Science

Editor's Note

Prof. Dr. Md. Abdus Samad

----

Analyzing Infant Mortality in Rural Bangladesh: A Frailty Modeling Approach

Bikash Pal

In practice, it may happen that data may arise from a hierarchical structure i.e., a cluster is nested within another cluster. In this case, nested frailty model is appropriate to analyze survival data to obtain optimal estimates of the parameters of interest. To identify significant determinants of infant mortality in rural Bangladesh, survival data have been extracted from Bangladesh Demographic and Health Survey (BDHS), 2014. Because of the presence of two-level clustering in data, nested frailty model has been employed for the purpose of analysis. Recommendations have been suggested based on the results obtained from the survival model to reduce the infant mortality in rural Bangladesh to a great extent.

Facial Expression Identification using Regularized Supervised Distance Preserving Projection

Sohana Jahan

Facial expression recognition is one of the most reliable and a key technology of advanced human-computer interaction with the rapid development of computer vision and artificial intelligence. Nowadays, there has been a growing interest in improving expression recognition techniques. In most of the cases, automatic recognition system’s efficiency depends on the represented facial expression feature. Even the best classifier may fail to achieve a good recognition rate if inadequate features are provided. Therefore, feature extraction is a crucial step of the facial expression recognition process. In this paper, we have used Regularized Supervised Distance Preserving Projection for extracting the best features of the images. Numerical experiment shows that the use of this technique outperforms many of state of art approaches in terms of recognition rate.

Synthesis and Characterization of Mixed Ligand Complexes of Copper(II) with Adenine and Dicarboxylic Acids

Sumaia Abdullah

Three mixed ligand complexes of copper(II) with adenine and dicarboxylic acids have been synthesized. The resulting complexes
were characterized by their melting point, solubility, metal content analysis, FT-IR and UV-visible spectroscopy, magnetic
measurement, thermal analysis, cyclic voltammetric measurement and X-ray powder diffraction study. The products are
microcrystalline powder, slightly soluble in water and decompose at high temperature. Under experimental condition, the ligands
adenine (Ade) behaves as a neutral ligand, whereas oxalic acid (OxH2), succinic acid (SucH2) and tartaric acid (TarH2) are doubly
deprotonated to form dianionic ligands that are coordinated to the Cu(II) ion. The Cu(II) content analysis of the complexes confine
to their stoichiometry [Cu(Ade)(L)(H2O)] (L = Ox, Suc, or Tar dianion). Electrochemical redox behavior of the complexes in their
reaction medium was also examined. They exhibit quasi-reversible one-electron transfer processes.

Pregnancy Complications in Women with PCOS: A Meta-Analysis

KM Tanvir

Polycystic ovary syndrome (PCOS) is a hormonal disorder which affects women in their reproductive ages. There might have association between maternal and neonatal complications with PCOS. The goal of this study is to look for the association between polycystic ovary syndromes (PCOS) and obstetric complications through meta-analysis on the basis of previously published studies from 2000 to 2019. In the current endeavour 17 studies involving 1975 women with PCOS and 10812 controls were selected for pooling. During pregnancy period women who have PCOS showed a significantly higher risk of developing gestational diabetes mellitus (GDM) (RR 2.51; 95% CI: 1.71-3.65), pregnancy induced hypertension (PIH) (RR 2.76; 95% CI: 2.04-3.72), preeclampsia (RR 2.14; 95% CI: 1.34-3.40), preterm birth (RR 1.41; 95% CI: 1.08-1.85) compared to controls. Neonatal birth weight is significantly lower (SMD -0.18; 95% CI: -0.33 to -0.03) compared to controls. It is concluded that women who have PCOS have greater chance of having pregnancy complications as well as risk of neonatal complications such as having low birth weight.

Performance Enhancement of an a-Si:H/μc-Si:H Heterojunction p-i-n Solar Cell by Tuning the Device Parameters

Md. Nazmul Islam Sarkar

In this work, the solar cell design parameters like- layer thickness, bandgap, donor and acceptor concentrations are varied to find optimum structure of a hydrogenated amorphous silicon (a-Si:H) and hydrogenated microcrystalline silicon (μc-Si:H) heterojunction p-i-n solar cell. A thin a-Si:H p-layer of 1 to 5 nm followed by a thick a-Si:H i-layer of thickness 1400 to 1600 nm and then thin n-layer of thickness 1 to 5 nm with acceptor concentration of 102 cm−3 and donor concentration of 1020 cm−3 and the bandgaps of p-, i-, and n- layers with higher bandgaps closer to 2.2 eV for a-Si:H p-layer, 1.85 eV for a-Si:H i-layer, and 1.2 eV for μc-Si:H n-layer have showed better performances. The optimum cell has a JSC of 18.93 mA/cm2, VOC of 1095 mV, Fill factor of 0.7124, and efficiency of 14.77%. The overall external quantum efficiency of the numerically designed cell also remained very high from 85-95 % for wavelengths of 300-650 nm range. This indicates that the device will perform its best under both high and low frequency i.e. ultra-violet, near visible and visible light wavelengths.

Studying the Third Cumulant of the Mixture of Dirichlet-multinomial Distributions

Farzana Afroz

Traditionally, the overdispersion parameter is estimated by using Pearson’s lack of fit statistic or the Deviance statistic , which do not perform well in the case of sparse data. This paper particularly focuses on an estimator of overdispersion parameter which was proposed for sparse multinomial data. The estimator was derived on the basis of an assumption on the 3rd cumulant of the response variable.When the data comes from the Dirichlet-multinomial distribution is known to have the lowest root mean squared error comparing to the other three estimators. In this paper the 1st to 3rd order raw moments of the finite mixture of Dirichlet-multinomial distributions are derived, which results in complicated mathematical expressions. Furthermore, it is found that the 3rd cumulant of this mixture does not satisfy the assumption which is considered in the derivation of .

Track and Landfall Characteristics of Very Severe Cyclonic Storm Fani over the Bay of Bengal using WRF Model

Md. Shakil Hossain

An attempt has been made to assess the capability of the Weather Research and Forecasting (WRF) model in simulating the track and landfall characteristics of Tropical Cyclone (TC) Fani (25th April – 05th May 2019) over the Bay of Bengal (BoB). WRF model has conducted on a single domain of 10 km horizontal resolution using Global Data Assimilation System (GDAS) data (0.25×0.25). The model predicted outcomes show auspicious agreement with the observed datasets of the Bangladesh Meteorological Department (BMD) and India Meteorological Department (IMD). It is found that the diminished lead time of the model run plays a crucial role in delivering good consistency with the minimum forecast uncertainty. A strong correlation between the track and intensity forecast deviations has also been determined. According to the results, the model simulation which captures the minimum deviation in the intensity forecast also ensures better track prediction of the system. The feasibility of the track and landfall forecast by the model even up to 27 hr advance is reasonably well. Finally, it can be decided that the model is capable to predict the cyclonic storm Fani precisely and it can be chosen confidently for future events over the BoB.

The Impact of Variable Fluid Properties on Biomagnetic Maxwell Fluids Past a Stretching Sheet with Slip Velocity and Heat Generation/Absorption

M. G. Murtaza

The impact of variable fluid properties (viscosity and thermal conductivity) and magnetic dipole on biomagnetic Maxwell fluid past
a stretching sheet with slip velocity and heat generation/absorptionhave been studied. Similarity transformation technique is
adopted to obtain the self-similar coupled nonlinear ordinary differential equations. Using similarity variable, the basic governing
equations with boundary conditions are transformed and solved in bvp4c technique with MATLAB software. The contribution of
different pertinent parameters such as viscosity, thermal conductivity and ferromagnetic parameter on the flow profiles with
physical quantities are analyzed and examined through graphically. Results shown that with increasing ferromagnetic parameter,
slip parameter, Maxwell parameter, velocity decreases but temperature increases. For accuracy of the proposed model to compare
our numerical results in numerically and graphically with the previousliterature under some limiting cases and a good agreement is
found.

Numerical Method for 2D Quasi-linear Hyperbolic Equation on an Irrational Domain: Application to Telegraphic Equation

Bishnu Pada Ghosh

We develop a novel three-level compact method (implicit) of second order in time and space directions using unequal grid for the numerical solution of 2D quasi-linear hyperbolic partial differential equations on an irrational domain. The stability analysis of the model problem for unequal mesh is discussed and it is revealed that the developed scheme is unconditionally stable for the Telegraphic equation. For linear difference equations on an irrational domain, the alternating direction implicit method is discussed. The projected technique is scrutinized on several physical problems on an irrational domain to exhibitthe accuracy and effectiveness of the suggested method.