In this paper, we propose efficient Bayesian methods to analyze longitudinal ordinal data with missing values using multivariate probit models. Longitudinal ordinal data with substantial missing values are ubiquitous in many scientific fields. Specifically, we develop the Markov chain Monte Carlo (MCMC) sampling methods based on the non-identifiable multivariate probit models and further compare their performance with the one based on the identifiable multivariate probit models. We carried out our investigation through simulation studies, which show that the proposed methods can handle substantial missing values and the method with marginalizing the redundant parameters based on the non-identifiable model outperforms the others in the mixing and convergences of the MCMC sampling components. We then present an application using data from the Russia Longitudinal Monitoring Survey-Higher School of Economics (RLMS-HSE).
In this study we assess the extent to which teaching and learning standards meet the criteria set by the education evaluation committee in Jordan. The viewpoints of students and teachers at Yarmouk University were considered. To fulfill the investigation's aims, the researcher employed a descriptive study design and utilized a quantitative technique of three domains: (attributes about graduates and educational results, curriculum, and the quality of instruction and student assessment), using the questionnaire as the primary instrument for data collection. The research involved a sample of (501) students and faculty members, comprising both male and female individuals from Yarmouk University. The participants were chosen through a random technique. The study utilized an independent sample t-test to assess the statistical significance of the effects of gender and participation position on the criteria for learning and teaching. The study's findings indicate that several elements, such as compliance with graduate qualities, educational outcomes, curriculum design, and the quality of instruction and assessment standards, influence the standards of teaching and learning. The results demonstrate that there is no statistically significant variation in the levels of teaching and learning standards when accounting for gender and participation position factors.
Model fitting and risk estimation are an important everyday aspect of a successful financial institution. Consequently, in this paper, we discuss risk quantification of the South African Industrial Index (also known as J520) using 22 standard light- and heavy-tailed statistical distributions. Given the importance of the J520 index (since it has the highest market capitalization in the Johannesburg Stock Exchange), investors may have a very keen interest in fully understanding the loss and gain returns characteristics and underlying statistical distribution’s properties, including tail properties. Thus, an in-depth goodness-of-fit evaluation is conducted by assessing six different tests (i.e. Kolmogorov-Smirnov, Anderson- Darling, Cramer von Mises, negative log-likelihood, Akaike information criterion, Bayesian information criterion) as well as two risk measures (i.e., value-at-risk and tail value-at-risk) are computed and interpreted within the context of J520 index. It is observed that the best distribution to fit to loss returns are the inverse Burr or transformed beta while for the gain returns it is either transformed gamma, inverse Burr or generalized beta distributions. The latter distributions strike a better balance with respect to excellent goodness-of-fit and risk measures that are very close to the corresponding ones for the empirical distribution. Given our findings, it may not be advisable for investors to hold very long positions in the J520 index since loss returns have much higher leptokurtic and heavy-tailed as compared to the lighter-tailed gain returns. Therefore, the growth shares observed over the long term indicate that a more prudent strategy would be to consider shorting the index. By doing so, investors could better align their strategies with the highly likely potential for substantial drawdowns inherent in this market.
Financial forecasting using deep knowledge and linear regression methods is explored in this study, which looks at the association between financial metrics and bank presentation from 2003 to 2012. By analyzing a complete dataset, the research identifies macroeconomic variables such as interest rates and rise, together with crucial success drivers such as GDP growth, leverage, and liquidness ratios. We can investigate linear and non-linear connections by mixing several mechanism learning models, such as ARIMA and Random Woodland. However, different banks' risk profiles and strategic prospects can be better understood through situation analysis and PCA-based clustering. The findings prove that rising GDP is a robust measure of economic wealth, but inflation cuts into pays. Institutions that are vulnerable to interest rate variations benefit from interest rates. Gathering shows that various banks adopt separate financial strategies, and scenario analysis shows that monetary results are very sensitive to changes in influence. By faithfully depicting financial doubt, the proposed models demonstrate pliability even during economically unstable times like the 2008 monetary crisis. This research emphasizes the importance of proactive risk organization and specialized forecasting tactics; upcoming studies should use hybrid modeling methodologies and join more macroeconomic variables. In the ever-changing world of finance, these advances aim to improve the correctness of predictions and planned decision-making.
In this paper, we explore the impact of price misrepresentation on pre-emption cases. It examines the legal and practical challenges faced when claiming recovery based on inaccurate prices and discusses the appropriate legal means to challenge forgery. The findings highlight the pre-emptor's right to challenge price falsehood in cases of increased prices and the absence of a specific legal method for filing an appeal. The study recommends clarifying the requirements for challenging price falsity and providing acceptable forms of evidence. Finally, it contributes to a better understanding of the topic and suggests improvements in the process.
This paper proposes a new life distribution with both unbounded and bounded support. The proposed Extended Exponentiated Lomax (EEtLx) distribution is derived from the New Extended Exponentiated-G (NEET-G) family of Elgarhy, Haq, Gozel, and Nasir (2017). The maximum likelihood estimation method is used to estimate parameters of the proposed distribution and derive the relevant properties. The model is applied to patient relief times (minutes) after receiving a particular analgesic using different selection criteria (Log likelihood, Akaike Information Criteria (AIC), and Bayesian Information Criteria (BIC)). Resulting distributions are compared to well-established lifetime distributions in literature. The total time on test (TTT) plot is also used to determine whether the hazard rate function for this data increases or decreases, which subsequently reveals a decreasing trend. The proposed distribution remains the best fit compared to competing distributions. It is recommended that this distribution can be applied, not only in medical science, but also in reliability science, engineering, and economics fields.
This article presents a new extension of the three-parameter log-logistic distribution, the so-called heavy-tailed log-logistic (HTLL) distribution. Some important mathematical properties of the new HTLL distribution are calculated. In addition, some numerical results of moments for the HTLL distribution are calculated. Extensive simulations were performed to investigate the estimation of the model parameters using many established approaches, including maximum likelihood (ML), least squares (LS), weighted least squares (WLS), Cramer-von Mises (CVM), Anderson-Darling (AD), and right-tail Anderson-Darling (RTAD). The simulation results show that the AD approach has the highest efficiency among these approaches. The usefulness of the newly proposed model is demonstrated by analyzing two real data sets.
Nonparametric kernel estimates used in this work aim to compare different treatment options by examining the recorded medical data. The following use of the suggested strategy depends on a kernel function and a parameter called bandwidth. The Nadaraya-Watson kernel (NWK) estimation is a necessary nonparametric kernel estimator used in regression models. A new Nadaraya-Watson regression estimate depends on the hyperbolic secant kernel (HSK) with fixed bandwidth (FNW) and Variable Bandwidth (VNW) is proposed.. We calculated some properties of the unknown regression function estimator, including bias, variance, optimal bandwidth, and a global measure of error criterion mean square error. Finally, simulation and three real data sets are used to evaluate its performance. Results from simulation and real data showed that the VNW using HSK is more effective than the FNW based on Average Mean Square Error Criterion. Also, Nadaraya-Watson using HSK function is more effective than Nadaraya-Watson using the Gaussian kernel density function.
The system under consideration has 𝑛 independent components, and its operation relies on the functioning of at least 𝑘 components, 1 ≤ 𝑘 ≤ 𝑛. The system experiences (𝑛 + 1) distinct shocks. Shock 𝑗 impacts the 𝑗!" component, 𝑗 = 1, 2, . . . , 𝑛, while shock (𝑛 + 1) simultaneously impacts all components. Any shock is lethal if its magnitude is below or above the component-designed thresholds 𝑑# or 𝑑$, respectively. A shock is characterized by its magnitude and arrival time, forming a bivariate random vector. The bivariate random vectors specifying the magnitudes and the arrival times of the shocks are assumed to be independent and follow non-identical bivariate distributions. The reliability of a 𝑘-out-of-𝑛: 𝐺 system under the influence of this type of shocks is derived. The reliability of parallel and series systems is obtained as special situations. The bivariate Pareto type I distribution is applied as an example of the bivariate distribution of the magnitude and arrival time of the shocks. Furthermore, numerical illustrations are conducted to highlight the theoretical results obtained.
ThispaperfocusesonthedevelopmentofapproximateBayesestimatorsfortheshapeparametersofthegeneralizedinverted Kumaraswamy (GIKum) distribution. The estimators are based on a progressive first-failure censored plan. The study considers both maximum likelihood and Bayesian estimations using a gamma-informative prior distribution for the parameters, as well as the reliability function, hazard rate, and reversed hazard rate functions. To obtain the estimators, the paper employs Lindley’s approximation and utilizes Markov Chain Monte Carlo (MCMC) methods. The Bayes estimators are derived with respect to both symmetric (squared error) and asymmetric (linex and general entropy) loss functions. In order to assess the performance of the proposed estimators, the paper presents numerical results obtained through a simulation study involving different sample sizes.
Measuring the impact of a binary treatment on a binary response variable is of great interest in many medical, social and economic applications. Estimating such effect is very important when the endogeneity problem occurs. This research proposes a bivariate logit model to control endogeneity when the structural errors of the two equations are correlated. The copula approach will be applied to estimate the dependence between the binary treatment and the binary response; and hence, the joint normality assumption of the structural error is irrelevant. For estimation, the maximum likelihood method will be applied to estimate the model parameters. The performance of the copula bivariate logit model in estimating the dependence between the binary treatment variable and the binary response variable is assessed by the Average Treatment Effect (ATE) criterion in both simulation study and real medical data.
This paper aims to investigate the implications of divergent thinking (DT) on the 'Habit-Bound Thinking' of undergraduate design students. The goal of this research is to find a way to develop creative designs that accurately mimic 2D frequent shapes. The results of this semi-experimental study gathered and analyzed observational data to evaluate the participants' divergent thinking abilities to generate a wide range of ideas or solutions to a given situation. Two colleges' worth of 120 students were split evenly between the control and experimental groups. This data set addressed the study's primary research question and produced substantial changes in the outcomes. Students in the control group improved significantly on tests measuring component knowledge and creative ability after receiving training and creating the graphical models. These results back up the theory that student creativity is fostered via project-based learning. The establishment of workshops led to a rise in originality
Lomax distribution has been studied by many statisticians due to its important role in reliability modeling and lifetime testing. The bivariate Lomax distribution is an important lifetime distribution in survival analysis. In this article, a bivariate Lomax distribution is constructed based on Clayton copula. The joint probability density function and the joint cumulative distribution function are derived in closed forms. Some properties of this bivariate distribution are discussed. The maximum likelihood and Bayes estimators for the unknown parameters are derived. Also, the maximum likelihood and Bayesian two-sample prediction for the future observations are obtained. The performance of the proposed bivariate distribution is examined using a simulation study. Finally, one real data set under the proposed distribution is considered to illustrate its flexibility and applicability for real-life applications.
The widespread use of fossil fuels for global energy production significantly contributes to global warming. This study presents a comparative analysis of various machine learning models, which are the long short-term memory (LSTM) network, support vector regression (SVR), and gradient boosting method (GBM). Gaussian process regression (GPR) is a benchmark model across different forecasting horizons. The study uses South African wind speed data from 1 January 2018 to 31 December 2021, sourced from the Western Cape province. The dataset underwent preprocessing, and diverse feature selection techniques were implemented to enhance model accuracy. Performance evaluation of the models was done using mean absolute error (MAE), root mean squared error (RMSE), and mean absolute scaled error (MASE). Results indicate that SVR exhibits superior accuracy to other models for two distinct forecast horizons (h = 670 and h = 1339), respectively. Additionally, GPR surpasses other models for the forecasting horizon h = 224. This study provides insights into the comparative strengths and weaknesses of different machine learning models for wind speed prediction, which could be useful in selecting an appropriate model for future applications in renewable energy and weather forecasting. Potential areas for future research include improving prediction accuracy via ensemble deep learning algorithms and incorporating additional meteorological variables. Moreover, investigating temporal dynamics, broadening geographical coverage and integrating uncertainty quantification methods can improve wind speed prediction, thereby facilitating more effective renewable energy planning and decision-making processes
A number of research settings involve data having a multilevel (hierarchical) structure. Failing to take into account such hierarchical structures may lead to wrong statistical inference. The objective of this study was to explore the correlates of employment status among graduates of Ethiopian Higher Education Institutions nested within their fields of specialization using hierarchical two‐level logistic regression model. The study sample consisted of 2569 graduates nested within 20 fields of specialization. The impact of level-1 covariates and fields of specialization on employment outcome was explored using average marginal effects. The estimated variance of the random (field of specialization) effects was found to be significant – an indication that a multilevel model is appropriate. Likelihood ratio test confirmed that the multilevel random coefficients model was a better fit to the data. Besides demographic and socio-economic characteristics of graduates (and their households), the dimensions related to the skills and competences of graduates, namely perceived level of training of graduates on technical/practical skills and soft skills at their universities as well as cumulative grade point average, had significant influence on employment outcomes. The results also revealed that the effect of practical skills on graduate employment varied across fields of specialization. Average marginal effects analysis indicated that the probability of employment was lower for graduates who were unmarried, aged below 25, with disability, from a low income family and those graduates who perceived that their preparation for practical as well as soft skills required in the job market was less than expected/very poor. Moreover, practical and soft skills differentials in the probability of graduate employment were found to diminish with an increase in cumulative grade point average. To improve graduate employment, provision of adequate training in practical/technical and soft skills is recommended.
In this paper, we examine the impact of the pandemic on university students in Jordan, focusing on their digital citizenship and attitudes towards distance learning. A cross-sectional study was conducted using an online survey, with 780 participants. The findings reveal that distance learning is widely applicable, with 89.87% of students utilizing live online learning content. The majority of students exhibit responsible online behavior, respecting others and refraining from bullying. They also actively engage in digital activities, express their opinions, and share their expertise online. Additionally, many students use technology for legitimate purchases during the pandemic. However, 95% of participants are aware of the potential health issues caused by excessive digital device use. Finally, it is shown that the university students in Jordan demonstrate good knowledge and positive attitudes towards digital citizenship, highlighting the importance of digital literacy and responsible online behavior in the face of the COVID-19 crisis.
In this paper, a parallel-series system is improved. All components assume independent and identically distributed. The Lindley distribution with three parameters is assumed to be a lifetime distribution for the components. Four methods are used to improve the performance of the parallel-series system. The γ-fractiles and equivalence factors are derived. Finally, numerical results are discussed.
: This paper proposes using maximum entropy approach to estimate the parameters of the Kumaraswamy distribution subject to moment constraints. Kumaraswamy [7] introduced the double pounded probability density function which was originally used to model hydrological phenomena. It was mentioned that this probability density function is applicable to bounded natural phenomena which have values on two sides. The distribution share several properties with the beta distribution and it has the extra advantages that is possesses a closed form distribution function, but it remained unknown to most statisticians until it was developed by Jones [6] as a beta-type distribution with some tractability advantages in particular as it has fairly simple quantile function and it has explicit formula for L-Moment. Using the principle of maximum entropy to propose new estimators for the Kumaraswamy parameters and compared with maximum likelihood and Bayesian estimation methods. A simulation study is performed to investigate the performance of the estimators in terms of their mean square errors and their efficiency.
In this article, with the use of the Bayesian, Expected Bayesian, and H-Bayesian estimation methods, the shape parameter and hazard rate of the Kumaraswamy distribution are calculated. Three separate loss functions the Squared Error loss function(SELF), the Precautionary loss function(PLF), and the Asymmetry loss function(ALF) along with an informative prior the Gamma prior are used to provide the estimates . The definitions and properties of the proposed estimators are given. Monte Carlo simulation is used to compare all of the estimates in terms of mean square error (MSE). Taking data from the real world, different estimation methodologies’ efficacy has also been studied. Numerical analyses show that the E-Bayesian estimates outperform the Bayesian and Hierarchical estimates.
The climate change crisis is negatively affecting the world and is the focus of many researchers attention for its life-threatening economic and climate impact on Earth. Therefore, this study aims to estimate the joint distribution function (EFXY) of both daily solar radiation (S) and daily maximum temperature (T) along with the Markov property. In this study, three-parameter distributions have been utilized with S and T, which are generalized extreme value (GEV) and Weibull (W-3P), respectively. Each of these parameters and the joint distribution function (𝐹(𝑋, 𝑌)) have been estimated. Four real data of S and T in Queensland, Australia during two consecutive years are applied. The method of maximum likelihood estimation (MLE) is applied on the proposed distributions of S and T to estimate their parameters, which was validated using Goodness-of-Fit tests. In addition, the logarithmic (LFXY) model and the multi-regression model (MFXY) for 𝐹(𝑋, 𝑌) are obtained. The results have been compared and the EFXY and LFXY are found to be non-equivalently, while the EFXY and MFXY are equivalent and homogeneous, confirming the validity of the joint distribution function estimate with the least error. Thus, the climate change probabilities are more accurately predictable by knowing both X and Y or by knowing both 𝐹(𝑋) and 𝐹(𝑌) with minimal error.
