Royal Statistical Society Nigeria Local Group Annual Conference Proceedings

Sampling Minoritised Populations: Experiences from the Evidence for Equality National Survey on the Impact of COVID-19 on Ethnic and Religious Minority Groups in Britain

Natalie Shlomo — 2025-04-07

The rapid improvements in our ability to conduct fast and cost-effective online surveys, together with advances in statistical theory to adjust for selection biases in nonprobability sampling, has led to opportunities to consider alternative sampling approaches for hard-to-capture minoritised populations. In this paper, we present an application of a carefully designed nonprobability online web survey to capture the experiences of ethnic and religious minority groups in Britain of the Covid-19 pandemic. The survey, The Evidence for Equality National Survey (EVENS), was funded by the Economic and Social Research Council (ESRC) in the United Kingdom and implemented by the Centre on the Dynamics of Ethnicity (CoDE) and Ipsos (a Survey Organisation). We describe here the robust design of EVENS, the data collection monitoring using quotas and facilitated by a responsive survey design framework including estimating Representativity (R-) Indicators. We also discuss the survey adjustment weighting approach that we followed to mitigate for selection biases. We conclude with lessons learnt and further recommendations for sampling minoritised populations.

THEME:POPULATION DATA FOR SUSTAINABLE SOCIO-ECONOMIC DEVELOPMENT

Prof. O.M. Olayiwola — 2025-04-07

NABLE SOCIO-ECONOMIC DEVELOPMENT

DATE: MARCH 5 - 6, 2025

THEME: THE POWER OF STATISTICS FOR A BETTER FUTURE

SPONSORED BY IASS AND RSS (UK)

ORGANIZING COMMITTEE

MODELLING MONTHLY WIND SPEED IN NORTH WEST NIGERIA: ExAR-FIGARCH AND ExAR-GARCH COMPARATIVE ANALYSIS

Shukurana Shehu — 2025-04-07

Understanding wind speed is the key planning renewable energy projects studying climate patterns and forecasting weather. In our study we explore how monthly wind speeds behave in North West Nigeria using an advanced model known as the Exponential Autoregressive-Fractional Integrated Generalized Autoregressive Conditional Heteroscedasticity (ExAR-FIGARCH). This model not only captures the lingering effects of past wind speeds but also accounts for unpredictable shifts over time. To see how its stacks up, we compare its performance against the more traditional ExAR-GARCH model, which mainly focuses on short-term fluctuations. We estimated two models and results showed the ExAR-FIGARCH model is better based on serial correlation analysis, efficient parameters and measures of accuracy, along with their ability to forecast future values. Our findings suggest that embracing long-memory effects in wind speed analysis could provide better insights into the region’s wind energy potential.

Key words: Wind speed, ExpAR-FIGARCH, ExpAR-GARCH model Long Memory, Volatility modelling.

RISK FACTOR OF VESICOVAGINAL FISTULA IN KEBBI STATE

Tolulope O. James — 2025-04-07

Vesico Vaginal Fistula (VVF) is still a major problem in developing countries of the world, especially in North Western Nigeria. This study examined the risk factors that affect the odds of having VVF. A sample of three hundred (300) questionnaires were used to collect data from some local government in Kebbi State, Nigeria. Logistic Regression model was used to predict the odd ratio or probability of being affected, to ascertain the risk factors of Vesico Vaginal Fistula. The result showed that factors such as previous VVF, X-ray of pelvis, injecting through veins, and physical examination of the vagina significantly increased VVF odds. Conversely, knowledge of VVF causes and delivering in a hospital reduced VVF odds, emphasizing the role of awareness and healthcare accessibility. While certain Socio-Demographic factors showed no significant association. The study recommends targeted measures, and the state government should create more enlighten atmosphere for the people about VVF and its consequences

Keywords: Vesico Vaginal Fistula (VVF), risk factor, logistic regression, odd ratio

FORECASTING INTERNALLY GENERATED REVENUE OF KADUNA STATE USING ARFIMA MODEL

Muhammad Idris Usman — 2025-04-07

Accurate forecasting of internally generated revenue (IGR) is crucial for effective fiscal planning and sustainable economic development. This study applies the Autoregressive Fractionally Integrated Moving Average (ARFIMA) model to forecast the IGR of Kaduna State, Nigeria. ARFIMA is particularly useful for modeling long-memory processes, which are common in financial and economic time series. The data used for the study was obtained secondarily from Kaduna State Internal Revenue Service (KADIRS). The stationarity of the data was assessed using Augmented Dickey Fuller (ADF) and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) tests. The long memory parameter d of the ARFIMA model was estimated using the Geweke and Porter-Hudak (GPH) method. The presence of a long memory structure was revealed by the sample autocorrelation function. Based on the information selection criteria, using AIC, BIC, and HQC, two optimal time series models were selected. But the prediction power of ARFIMA (3,0.423636,4) model is better and suitable for monthly periods forecasting, as such the model best fit the data. Thus, the findings can be used to provide accurate and reliable forecast of Kaduna State IGR for better revenue planning and economic policy formulation.

Keywords: ARFIMA model, forecasting, internally generated revenue, long-memory.

Sensitivity analysis on joint modelling of longitudinal and mixture cure outcomes with data missingness and outliers using INLA: application to aortic valve replacement surgery data

A. H Ekong — 2025-04-08

Joint modelling under Bayesian paradigm has gained a lot of traction especially with sampling-based estimation, however approximate Bayesian estimation of integrated Laplace approximation (INLA) is slowly gaining grounds. Prior specification has also been a recurring discuss in Bayesian analysis with prior sentivity becoming part of the data analysis process. This work presents joint modelling of longitudinal and cure proportion using latent Gaussian model with INLA and prior sensitivity analysis for the model in the presence of data value missingness and outliers. The approach assumed inverse-Wishart prior distribution for the covariance matrix of the random effects and Gaussian priors for the joint model fixed effects, while the penalised complexity prior was assumed for the Weibull shape parameters of the baseline hazard function. Four different prior specification settings were studied for fixed and random effects and the association parameter. The study was applied to aortic valve replacement surgery data to assess the effects of covariates on a biomarker and risk of event, with spline trajectories. The best prior setting was arrived at via the lowest values of DIC, WAIC and log marginal-likelihood and was Gaussian prior for fixed effects and association parameter each with (mean, precision) values as (0, 0.001), (0, 0.001), (0, 0.001), and parameters from Wishart distribution on the precision matrix for random effects as (100, 1) and it gave robust results with missing values and outliers. The posterior estimates from the best prior settings showed significant covariates on the biomarker and on the conditional failure time latency model. The study contributes to the literature on approximate Bayesian alternative to jointly modelling of longitudinal and mixture cure outcomes in the area of prior specification and data value missingness and outiers.

Keywords: prior specification, association structure, Laplace approximation, shared random effect, nonlinear trajectory

IMPACT OF SAMPLE SIZE ON THE ACCURACY OF PARAMETER ESTIMATION IN VARIOUS PROBABILITY DISTRIBUTIONS

Modu Kaka — 2025-04-08

Despite the general consensus that larger samples improve estimation accuracy, there was limited comprehensive understanding of how this relationship differed among various distributions. This study filled this gap by systematically analyzing the effect of sample size on parameter estimation accuracy. The objectives included evaluating the relationship between sample size and estimation accuracy for different distributions, comparing the performance of different estimation methods, identifying minimum sample sizes required for specified accuracy levels, and providing practical guidelines for researchers. Focusing on normal, binomial, Poisson, exponential, and gamma distributions, the study examined sample sizes ranging from small (n=10) to large (n=1000). The methodology included simulation studies to generate datasets, accuracy assessment using bias, mean squared error (MSE), and confidence intervals, and comparative analysis to identify patterns and trends. The expected outcomes included a detailed understanding of sample size effects on estimation accuracy, identification of minimum sample sizes for accurate estimation, and development of practical guidelines to enhance the efficiency and reliability of statistical analyses across various fields.

MODELLING AIRQUALITY OF ILORIN, KWARA STATE WITH FAMILY OF XLINDLEY DISTRIBUTION

Elijah Faniyi Toyin — 2025-04-08

Ozone, a major component of smog, forms through chemical reactions involving pollutants like volatile organic compounds and nitrogen oxides, primarily from vehicle emissions and industrial activities. As a key air quality indicator, high ozone levels pose environmental and health risks, causing respiratory issues such as asthma and bronchitis while increasing cardiovascular disease risk. This study proposes the Generalized XLindley Distribution to model ozone levels in Ilorin, Kwara State, and compares it with the Exponential-Lindley, Quasi XLindley, and Inversed XLindley distributions. The Kolmogorov-Smirnov test was used to assess goodness-of-fit, while parameter estimation was performed using the Maximum Likelihood Estimator and the Method of Moments. A simulation study explored model behavior across varying parameters, with survival and hazard functions analyzed for deeper insights. Model selection criteria, including AIC, AICC, BIC, and HQIC, were applied to evaluate efficiency. The Generalized Xlindley Distribution outperformed competing models, providing a more accurate fit. These findings enhance statistical modeling of air pollution data, improving air quality assessment and prediction in urban settings.

Keywords: Weighted Distribution, Xlindley distribution, Maximum Likelihood Estimation, Ozone, Airquality

INTRODUCTION

THE INVERSE LOMAX DISTRIBUTION WITH APPLICATIONS

Yunusa Falgore Jamilu — 2025-04-08

The Inverse Lomax Distribution, a flexible heavy-tailed distribution, has gained attention in statistical modeling, particularly in fields like finance, reliability engineering, and risk analysis. This research delves into the applications of the Inverse Lomax Distribution, offering insights into its robustness in various contexts. The parameters of the ILD were estimated using maximum likelihood method. Key advantages include its ability to model data with high skewness and heavy tails, making it a strong candidate for applications requiring robust outlier sensitivity. Additionally, the distribution’s straightforward form enables analytical tractability, facilitating parameter estimation and inference.

Keywords: Inverse Lomax Distribution; Inverse Lomax Log-Logistic Distribution; Skewed Distribution.

THE IMPACT OF CRUDE OIL PRICE SHOCKS ON THE NIGERIAN EXCHANGE RATE: A TIME SERIES ANALYSIS USING ARIMAX MODELS

R. O David, — 2025-04-08

Fluctuations in crude oil prices significantly impact the value of Nigeria's Naira relative to the US Dollar. Understanding the link between crude oil price shocks and exchange rate movements is vital for effective monetary policy and economic stability. This study examines the effect of crude oil price shocks on the USD/NGN exchange rate using time series analysis. Historical data on crude oil prices and the exchange rate were analyzed with ARIMAX models, incorporating crude oil prices as an exogenous variable. The ARIMAX (1,1,1) model provided the best fit based on AIC and BIC criteria, with diagnostic checks revealing no autocorrelation in residuals, ensuring model adequacy. The findings suggest that crude oil price changes exert a small negative effect on the exchange rate, but historical values and past errors play significant roles. The study concludes that while crude oil price changes influence the exchange rate, additional factors like inflation, foreign reserves, and government policies could have a greater effect. Recommendations include diversifying the economy and implementing policy reforms to improve exchange rate stability in Nigeria.

MODELLING RETURN UNPREDICTABILITY WITH THE ODD GENERALIZED EXPONENTIAL LAPLACE DISTRIBUTION: A SIMULATION STUDY

R. O David — 2025-04-08

The accurate modelling of return unpredictability remains a pivotal challenge in financial econometrics. Traditional models often assume a normal distribution for error terms, which fails to capture the leptokurtic and skewed nature of financial returns. This paper introduces the odd generalized exponential Laplace distribution (OGELAD) as an error distribution tailored for modelling asset return unpredictability. The proposed distribution addresses the limitations of conventional error distributions such as normal (NORM), skew normal (SNORM), normal inverse Gaussian (NIG), and skew generalized error distribution (SGED) in capturing key characteristics of financial returns, such as asymmetry and heavy tails. Using simulated data, the study evaluates the performance of the OGELAD within symmetric and asymmetric volatility models, demonstrating its effectiveness in modelling and forecasting return volatility. Diagnostic tests confirm that all error distributions, including the OGELAD, successfully eliminate ARCH effects from residuals, ensuring robust model performance. Notably, the positive and significant asymmetry parameter in the selected model highlights that positive shocks exert a smaller influence on volatility compared to negative shocks of the same magnitude. This finding underscores the relevance of the proposed distribution in capturing leverage effects observed in financial data. The OGELAD distribution consistently outperformed existing distributions in modelling and forecasting volatility, showcasing its potential for broader applications. It can be extended to multivariate settings for portfolio risk management and applied to high-frequency financial data to test its robustness under varying market conditions.

DEVELOPMENT OF LOGIT SKEWED EXPONENTIAL POWER (LSEP) DISTRIBUTION FOR MODELLING INFLATION RATE

Abayomi Olumuyiwa Ajayi — 2025-04-08

This study focuses on the development and application of the Logit Skewed Exponential Power (LSEP) distribution to model rates and proportion. The data considered were the Nigeria's inflation rate from 2008 to 2024. The study employs a systematic methodology that involves parameter estimation, graphical analysis using histograms, P-P plots, Q-Q plots, and model adequacy evaluation through information criteria such as Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). The estimated parameters for the LSEP model—Alpha (0.3001), Beta (1.2915), Miu (-0.7066), and Sigma (0.0197)—demonstrated its ability to address the skewness and scale of the inflation rate distribution. The Beta distribution's parameters—Alpha (1.367) and Beta (6.678)—revealed a simpler structure that was less capable of capturing the data's complexity. The evaluation of model adequacy through information criteria showed that the LSEP model had a lower log-likelihood value (66.26637) and lower AIC (-124.5327) and BIC (-115.9602) compared to the Beta distribution log-likelihood value (52.8371) and higher AIC (-101.6742) and BIC (-97.38794). The LSEP model’s parameter estimates indicate its robustness in addressing the complexities of the inflation rate, with lower AIC and BIC values compared to the Beta distribution.

EXPLORING THE PREDICTIVE POWER OF UNEMPLOYMENT RATES ON INFLATION IN EMERGING ECONOMIES IN AFRICA: A QUANTITATIVE APPROACH.

I K Adebayo — 2025-04-08

This study explores the predictive power of unemployment rates on inflation in emerging economies, with a focus on African regions. Using a quantitative approach, we analyzed data on unemployment and inflation rates across five distinct African regions: Eastern, Middle, Northern, Southern, and Western Africa. Descriptive statistics reveal notable regional disparities, with Southern Africa exhibiting the highest mean values for both unemployment (2.08) and inflation (1.90), while Western Africa showed lower inflation rates. The data analysis indicates that unemployment rates tend to have a negative skew and are somewhat platykurtic across regions, while inflation rates exhibit moderate negative skewness and leptokurtic distributions, suggesting more extreme values than expected. The ANOVA tests confirm significant differences in mean unemployment and inflation rates across the regions (p < 0.05). Further, regression analysis demonstrates a significant relationship between unemployment and inflation rates, with the log of unemployment rate positively predicting the log of inflation rate ( = 0.148, p = 0.000). However, the low R² value (0.016) suggests that while the relationship exists, it only explains a small portion of the variation in inflation rates. Overall, this study highlights regional variations in unemployment and inflation dynamics across Africa and provides empirical evidence of a positive though weak relationship between unemployment and inflation contributing valuable insights to economic policymaking in emerging economies.

Keywords: Unemployment rate, Inflation rate, Emerging economies, African regions, Quantitative analysis, Economic disparities.

EFFECT OF FUEL SUBSIDY REMOVAL ON THE SMALL – SCALE FARMERS IN IBARAPA CENTRAL LOCAL GOVERMENT, IGBOORA, OYO STATE

Kayode OLAKUNLE — 2025-04-08

The removal of fuel subsidies has been a contentious issue in Nigeria, with proponents arguing that it will lead to increase economy efficiency and opponents arguing that it leads to increase poverty and inequality. This study examined the impact of fuel subsidy removal on small – scale farmers in Ibarapa Central Local Government, Igboora, Oyo State. A survey of 100 small – scale farmers was conducted, and the data was analyzed using descriptive and inferential statistics. The results showed that the removal of fuel subsidies had a significant negative impact on the farming activities of small – scale farmers, leading to increased production cost, reduced agricultural production, and reduced income. The study recommended that policymakers should consider the potential impacts of fuel subsidy removal on small – scale farmers and the agricultural sector as a whole. Alternative energy sources, subsidies on agricultural inputs, and training and extension services were also recommended to mitigate the negative impacts of fuel subsidy removal on small – scale farmers.

Keywords:

Fuel subsidy removal, Small – scale farmers, agricultural production, income, Nigeria.

ODD TRANSMUTED RAYLEIGH- LOMAX DISTRIBUTION WITH ITS APPLICATIONS

J. Abdullahi — 2025-04-08

The flexibility and applicability of some known compound distributions in modelling and predicting diverse industrial datasets cannot be quantified. Nowadays, some of these existing distributions found it difficult in modelling different datasets most especially the ones that are highly skewed to either direction. On this note, this study focused on developing a distribution named Odd Transmuted Rayleigh- Lomax Distribution (OTRLD) which can capture the important data points. The cumulative distribution function (CDF), probability density function (PDF) and reliability function of the OTRLD were explicitly derived.
Consequently, the validity of this distribution has been checked and confirmed valid. The parameters of the new distribution were estimated using the technique of maximum likelihood. On the basis of distributional performance, the derived model was compared with existing distributions in the literature via real datasets. The results have shown that the proposed distributions outperformed their competitors.
Keywords: Odd Transmuted Rayleigh, Lomax, reliability, maximum likelihood.

DELAY PARAMETER OF TRANSITION VARIABLE IN THE SMOOTH TRANSITION AUTOREGRESSIVE (STAR) MODEL: IMPORTANT DETERMINANT

Asuquo Effiong Benjamin — 2025-04-08

This study investigates some determinant of appropriate delay parameter of the transition variable in the Smooth Transition Autoregressive (STAR) models, with emphasize on the model type and data characteristics through a refine method. Daily share price data were sourced from the Nigerian Exchange Limited, covering 10 years (January 2, 2014 to December 29, 2023), comprising 2,472 observations for each of the selected stocks: GTCO and STANBIC from the financial sector, and DANGCEM, BETAGLASS, and WAPCO from the industrial sector. Linearity tests demonstrated that DANGCEM, GTCO, and STANBIC share returns exhibit nonlinear characteristic of financial time series (FTS), whereas WAPCO returns remain linear and the most suitable delay parameter for each stock returns was determined. The Escribano-Jorda procedure was employed to select appropriate transition function. An asymmetric transition function was specified for DANGCEM and GTCO stock returns, while a symmetric transition function was identified for STANBIC and BETAGLASS returns. Consequently, asymmetric STAR models were fitted to DANGCEM and GTCO, and symmetric STAR models were fitted to BETAGLASS and STANBIC, using delay lengths within the range (1≤d≤p). The results indicated that the APLSTAR, LSTAR, and SPLSTAR models with the initially chosen delay lengths were optimal for DANGCEM, GTCO, and BETAGLASS stock returns, respectively. However, for STANBIC returns, the optimal SPLSTAR model was associated with a delay length different from the initially selected delay parameter. This study concludes that while the delay parameter of the transition variable is typically determined from the characteristics of the FTS, STAR model type also significantly influences the selection of an appropriate delay parameter.

COMPARATIVE STUDY OF THE ECONOMY OF NIGERIA BEFORE, DURING AND AFTER THE CORONA VIRUS DISEASE 2019 (COVID-19) LOCKDOWN

D. F. Nwosu — 2025-04-08

This study examines the Nigerian economy before, during, and after the COVID-19 lockdown by analyzing monthly data on six key economic indicators—Gross Domestic Product (GDP), inflation rate, unemployment rate, exchange rate, interest rate, and trade balance—from January 2019 to December 2021. Data were sourced from the National Bureau of Statistics (NBS), the International Monetary Fund (IMF), the Central Bank of Nigeria (CBN), and the World Bank. Repeated measures ANOVA and Friedman’s test (for non-normal data) were applied. Due to non-normality, the Friedman test was used to analyze GDP, interest rate, unemployment rate, and exchange rate, revealing significant differences across years. The Nemenyi post-hoc test identified statistically significant differences in GDP (2020–2021), interest rate (2019–2020, 2019–2021), unemployment rate (2019–2020, 2020–2021), and exchange rate (2019–2021). Repeated measures ANOVA was used for inflation rate and trade balance, showing significant differences in inflation rates but not in trade balance. Overall, five of the six economic indicators exhibited significant changes, underscoring the COVID-19 lockdown’s substantial impact on Nigeria’s economy.

Keywords: Corona Virus Disease 2019 (COVID-19), Economic Indicators, Friedman Non-Parametric Test, Nigeria’s Economy, Repeated Measures Analysis of Variance

MODELLING LONGITUDINAL AND SURVIVAL DATA WITH MULTIPLE BIOMARKERS

O. M Olayiwola — 2025-04-08

Background: Existing joint model for longitudinal and survival data captured both types of data, but there is dearth of information about methodologies that captured simultaneously the trajectories of multiple biomarkers over time. This study developed a joint statistical model that captured concurrent trajectories of multiple biomarkers using longitudinal and survival data. Data from the Mayo Clinic trial on Primary Biliary Cirrhosis was used to validate the model. The dataset comprised 424 patients that met eligibility criteria, with 312 actively participated in the trial. An additional 112 cases that participated not in the trial consented to basic measurements and survival monitoring

Methods: The joint model was developed by integrating a longitudinal sub-model (longitudinal outcomes over time) and a survival sub-model (the time until a specified event occurs) and compared with Mayor’s models. The longitudinal sub-model was represented by a linear mixed-effects model and the survival sub-model by the Cox proportional hazards model. The two sub models were connected using a shared random effect to capture the correlation between longitudinal trajectory and event risk. The model parameters were estimated using the Expectation-Maximization algorithm and diagnostic checks were carried out to validate the model.

Results: The results revealed consistent trends in serum bilirubin levels, significant differences in serum cholesterol between placebo and D-penicillamine groups, and gender-related disparities in survival outcomes. A 55% observed survival rate highlighted positive health outcomes, while an 8% incidence of liver transplants underscored the complexity of the targeted medical conditions. An even distribution of participants between interventions ensured a fair comparison, emphasizing the efficacy of D-penicillamine while acknowledging the challenging nature of the addressed health conditions. Gender-specific analyses showed significant associations, with females exhibiting a hazard of survival approximately 0.4913 times that of males.

Conclusion: The survival model identified significant associations between survival time and biochemical measurements with high predictive accuracy.

Keywords: Statistical Models, longitudinal data, survival data, biomarkers, clinical trial

COMPARATIVE ANALYSIS OF LINEAR AND NON-LINEAR REGRESSION MODELS ON IMPACT OF NIGERIAN POPULATION AND DEBT ON GROSS DOMESTIC PRODUCT (GDP)

K. O. Omosanya — 2025-04-08

Country’s population is a 2way adaptor, either beneficiary or detrimental depending on how the government channels it. But national debt is short term beneficiary and become economic trap if persistent in the long run on economic growth of the nation. There had been numerous studies that aimed at examining the effects of public debt and population growth on economic growth carried out over time across countries of the world, models of both linear and non-linear had been applied to solve problems of public debt and population on Gross Domestic product most of which are bereft of strategic empirical evidences suitability in developing countries. We consider Linear and Exponential growth model estimators for estimating the significant of population and National debt on the economic growth of Nigeria measured by GDP and realized that population growth rate has exerted significant positive impact on economic growth (GDP) While, National debt does not have any significant impact on economic growth. Moreover, AIC still remain the best model selection criterial and that linear regression model is recommended in fitting Nigeria economic growth on the population and public Debt.

Key Words: Akaike Information Criterial (AIC), Exponential growth rate, Gross Domestic Product (GDP), Linear growth rate, Population, Public Debt.

EXAMINING THE RELATIONSHIP BETWEEN MONETARY POLICY AND STANDARD OF LIVING IN NIGERIA: A REGRESSION ANALYSIS APPROACH.

Rokibat Adeola TIJANI — 2025-04-08

This study investigated the impact of monetary policy on the standard of living in Nigeria, using a multiple linear regression model with robust standard errors. The study employs annual report data from 2012 to 2021, sourced from the Central Bank of Nigeria. The result of the finding show that monetary policy has a significant negative impact on the standard of living, with 1% increase in monetary policy rate leading to a 0.25% decrease in the standard of living. It also finds that inflation rate, unemployment rate, GDP growth rate, exchange rate and government spending have significant impacts on the standard of living. The study controls for potential endogeneity using lagged values of the independent variables and includes interaction terms to examine nonlinear relationship. The findings of this study have important implications for monetary policy formulation and suggest that policymakers should consider the potential impact of monetary policy decisions on the standard of living in Nigeria.

Keywords

Monetary Policy, Standard of living, Multiple Linear Regression, Robust Standard Errors, Nigeria.

MODELING CHAOTIC BEHAVIOUR OF STOCK PRICE INDEX USING ITO STOCHASTIC DIFFERENTIAL EQUATION

O. O Okeji — 2025-04-08

Stochastic differential equation (SDE) have become an important tool for modeling the dynamics
of many random phenomena such as financial assets. In real applications, parameters of the
equation are unknown and need to be estimated and many times only discretely sampled data of
the process are available. Financial assets such as stock price are very chaotic and dynamic and
are often represented using stock price index, to reflect overall market sentiments and directions
of stock prices. Investing in stocks or equities is a speculative risk that is complex and
complicated to understand due to its chaotic behaviour. In this paper, attempt was made to study
this chaotic bahaviour via Ito SDE, the forward Kolmogorov equation (FKE). The parameters
estimation was done using Euler-Maruyama method. The model’s mean, variance and Akaike
Information Criterion (AIC) were obtained as 0.08, 896.56 and 4764.08 respectively, as against
ARIMA (1,0,0), (3,1,1) and (6,0,0) having AIC values of 5482.92, 5401.00 and 5433.50,
respectively. Hence the Ito SDE was better in describing stock price index and is therefore
recommended for practitioners and policy makers for sound decision making regarding stocks.
Keywords: Chaos; Diffusion process; Kolmogorov equation; nonlinear dynamics; stochastic
differential equation; Stock price index.

A NEW ROBUST METHOD OF DEALING WITH MULTICOLLINEARITY AND OUTLIERS IN REGRESSION ANALYSIS: SIMULATIONS AND APPLICATIONS

T. J Adejumo — 2025-04-08

Ordinary Least Squares (OLS) regression is known to be unreliable in the presence of outliers and multicollinearity, leading to biased parameter estimates, inflated standard errors, and reduced predictive accuracy. Limited studies have considered estimators that can co-handle the two problems (multicollinearity and outliers). However, to explore further methods, this study proposes the Robust M-version of the New Biased Based Estimator (M-NBB), designed to handle both multicollinearity and outliers effectively. Theoretical properties of the proposed estimator were established under fundamental conditions and validated through Monte Carlo simulations implemented in R-statistical programming. The simulation study involved 1,000 replications across eight sample sizes and three explanatory variables exhibiting varying degrees of multicollinearity. Additionally, 10% and 20% of the generated observations were contaminated with outliers of various magnitudes under error variances. The performance of the estimators was evaluated using the Mean Squared Error (MSE). The simulation results revealed that the proposed estimator outperformed existing estimators across all conditions. To further validate its effectiveness, the estimator was applied to real-life data. The findings suggest that the M-NBB estimator is a robust and reliable alternative for practitioners dealing with datasets affected by both multicollinearity and outliers.

Keywords: Ordinary Least Squares, Multicollinearity, outliers, Monte Carlo simulation, Estimator

A STUDY ON THE PERFORMANCE OF THE PARTIAL LEAST SQUARES REGRESSION IN HANDLING MULTICOLLINEARITY USING SIMULATED DATA

I. A Sadiq — 2025-04-09

Multicollinearity is a common issue in regression analysis which occurs due to the violation of the assumptions of regression that there is no correlation between the explanatory variables of the least square estimator, and because of the violation, the estimate of the parameters tends to be less precise and unreliable, and this leads to unstable inflated variance. Thus, the biased regression techniques which stabilize the variance of the parameter estimate were employed. This study focused majorly on the Partial Least Square Regression, a biased regression technique for overcoming multicollinearity, the strength and limitations of the method, and also the performance of the method when compared with the Principal Component Regression (PCR) using the Root Mean Square Error (RMSE) as a performance metric. A simulation study of data that follows a normal distribution with varying levels of multicollinearity was conducted to evaluate the accuracy, interpretability, and robustness of PLSR models and also in comparison to the PCR using the root mean square error (RMSE) as a performance metric. Based on this study, it is observed that the PLSR is more robust to multicollinearity than PCR, as it is less likely to produce unstable parameter estimates in highly correlated datasets. Therefore, this technique can be applied to the same distribution used in this study by varying the sample sizes. It can also be used to look at the behaviors of distributions other than those used in this study.

Keywords: Regression; Multicollinearity; Partial Least Square Regression; Principal Component Regression; Simulation

MODELLING REPEATED MEASURES DOSE-RESPONSE MORTALITY DATA USING GENERALIZED ESTIMATING EQUATIONS (GEE)

S. A. ADEH — 2025-04-09

ABSTRACT

Repeated measures dose-response mortality studies usually involve obtaining responses at different times on the same group of subjects, which often leads to correlation. A commonly used method for correlated dose-response mortality data is the Probit analytical technique which is suitable for data collected at one point in time and not for repeated measures. This study developed a Generalized Estimating Equations (GEE) using logistic regression for estimating the model parameters in repeated measures dose-response mortality data. The GEE model was applied to adult-termites mortality data observed at 6, 12, 18 and 24 hours respectively from an experiment conducted in the Entomology Division of the Nigerian Institute for Oil Palm Research, NIFOR, Edo State, Nigeria. In the experiment conducted, adult-termites were exposed to two plant extracts; Jatropha Curcas and Ricinus Cummunis at varying concentration levels (10%, 20% and 35%) respectively. The GEE estimated LT50 results for each plant extracts at varying concentration levels were given as J.Curcas (LT50=12.47hrs, 12.47hrs and 12.47hrs) and R.Cummunis (LT50=12.47hrs, 12.47hrs and 12.47hrs) which shows that the potency of the concentration levels is the same considering the time to mortality. Repeated measures logistic regression using GEE has proven to be a robust method in estimating LT50 since it consistently gave precise LT50 estimates with a smaller confidence interval, thus should be incorporated into studies of this nature as other existing methods for analyzing data from bioassay experiments.

Keywords: Repeated measures, dose-response, correlation, Probit analysis, GEE, Survival data, plant extracts, and mortality.

TRANSMUTED SINE GENERALIZED EXPONENTIAL DISTRIBUTION: DISTRIBUTIONAL PROPERTIES AND APPLICATIONS

Mustapha Dewu Muhammad — 2025-04-09

This research paper introduces a novel probability distribution called the Transmuted Sine Generalized-Exponential (T-SGEx) distribution, which is developed by combining the transmutation technique with the sine function and the Generalized-Exponential distribution. The distribution is designed to offer greater flexibility in modeling real-life data, particularly for datasets exhibiting complex behaviors such as skewness, heavy tails, or multimodal patterns. To understand the properties of the T-SGEx distribution, several distributional characteristics, including probability density function (PDF), cumulative distribution function (CDF), hazard rate function, moments, and quantile function, were examined. The study also explored the reliability properties of the distribution, which are critical for applications in survival analysis and engineering.

The estimation of the T-SGEx distribution parameters was performed using Maximum Likelihood Estimation (MLE). Real-life datasets from two previous studies were used to evaluate the performance of the T-SGEx distribution. The goodness-of-fit of the T-SGEx distribution was compared with two competing distributions: the New Weighted Exponential Distribution (NWED) and the standard Exponential Distribution (ED). Model comparison was conducted using the Akaike Information Criterion (AIC) and the negative log-likelihood (NLL). The results demonstrated that the T-SGEx distribution provides a superior fit to the datasets, as evidenced by lower AIC and NLL values compared to the NWE and Exponential distributions.

The findings of this study highlight the practical application of the T-SGEx distribution in modeling real-life data, particularly in fields such as reliability engineering, survival analysis, and environmental studies.

Keywords: Distribution, Quadratic rank transmutation map, Maximum likelihood estimation, Order Statistics, Transmuted Sine-G distribution

Predicting Patient Recovery Time Using Clinical and Lifestyle Variables: A Statistical Approach

I, K Adebayo — 2025-04-09

Understanding the factors influencing patient recovery time is essential for improving healthcare outcomes. This study applies a Multiple Linear Regression (MLR) model to predict recovery time using clinical (Age, BMI, Severity Score, Hospital Stay, Underlying Conditions, Medication Type) and lifestyle factors (Smoking, Alcohol Consumption, Physical Activity). A dataset covering a 10-year period (2014–2023) was analyzed. The MLR results reveal that Severity Score (β = 5.0021, p < 0.001) and Hospital Stay (β = 0.8052, p < 0.001) are the strongest predictors of recovery time. Lifestyle choices also significantly impact recovery: smoking (β = 1.5634, p < 0.001) and alcohol consumption (β = 1.4082, p < 0.001) extend recovery time, while higher physical activity (β = -0.9352, p = 0.0012) speeds it up. The model explains 98.5% of the variance (R² = 0.985) and has a low RMSE (1.91 days), indicating high accuracy. The findings highlight the need for personalized treatment plans that consider both medical conditions and lifestyle habits. Healthcare providers can use this model to predict recovery time more effectively and design interventions that encourage healthy lifestyle changes. Future research should explore non-linear effects and integrate additional biological markers for enhanced predictive accuracy.

Keyword: Clinical Variables, Healthcare, Lifestyle Factors, Multiple Linear Regression,

Predictive Modeling, Recovery Time

Assessing the Forecast Performance of EGARCH and Prophet Models in Analyzing the Effect of Holidays/Events on Nigeria Stock Price Returns under Normal and GED Assumptions of Innovations

U. Abdulaziz — 2025-04-09

Stock market price returns are greatly affected either positively or negatively by so many factors such as day-of-the-week, holidays/events (Childrens day, Valentine’s Day, Id el Kabir, Maulud, Christmas day, Eve day among others). And the effect of such factors on Nigeria Stock market was little or not been assessed and quantiﬁed. This paper aimed at studying the effects of holidays/events on the Nigeria Stock market price returns utilizing EGARCH and Prophet models under Normal and Generalized Error Distribution (GED) assumptions of innovations. The results of EGARCH model under Normal assumption of innovations revealed that the holidays that falls on Mondays and Fridays has signiﬁcant effects on the NSE returns; but the effects are not signiﬁcant on the other days. Similarly, the results of the Prophet model revealed that all the effect of holidays that falls on Mondays down to Fridays are not signiﬁcant. On the other hand, the EGARCH model (under GED assumption of innovations) and Prophet model agreed equally as all the holidays effects are not signiﬁcant in both the cases. With the help of MAE, RMSE, and MASE, EGARCH with GED assumptions of innovations performed better than EGARCH with Normal assumptions of innovations as well as the Prophet model.

Keywords—EGARCH, Facebook prophet, holidays/events effects, GED, NSE.

SIMULATION STUDY OF PARAMETERS’ INFLUENCE OF HESTON’S STOCHASTIC-JUMP MODEL ON VOLATILITY SMILE

Onyegbuchulem Chialuka Adline — 2025-04-09

Abstract
This paper aims at presenting Heston’s Stochastic-Jump model, then study the parameters influence of the model on volatility smile. Complete derivation of the Heston’s Stochastic-Jump model was presented. Simulation studies were conducted and results show that Heston’s Stochastic-Jump model addresses the shortcomings of the Black-Scholes because the way the volatility is modelled is more realistic from financial market’s point of view compared to the constant volatility assumption since it takes into consideration what is observed in financial markets. Hence, combining jumps and stochastic volatility therefore
produces models which are more flexible and that can accurately fit observable market data.

Keywords: Heston’s Stochastic Model, Heston’s Stochastic-Jump Model, Black-Scholes’ Model, Volatility Smile

ENHANCED POPULATION SIZE ESTIMATION USING ZELTERMAN-TYPE ZERO-TRUNCATED DISCRETE LINDLEY DISTRIBUTION UNDER ONE-INFLATED POISSON COUNT DATA

Nathan Samuel Agog — 2025-04-09

Capture-recapture analysis is widely used for population size estimation in various fields, including ecology, biology, social sciences and medicine. The Zelterman Poisson estimator obtained from Poisson distribution is commonly used for estimating population size in capture-recapture but it tends to underestimate counts when dealing with overdispersed data. To address this limitation, this paper proposes the Zelterman-type estimator (Zelterman-DLD) developed under Zero-Truncated Discrete Lindley (ZTDL) distribution for improved population size estimation. The paper evaluates the performance of two estimators; Zelterman-DLD and Zelterman Poisson estimator (Zelterman-POIS) using count data derived from a one-inflated Poisson distribution. These estimators were assessed across different population sizes under varying levels of one-inflation (10% and 20%). Variance estimation was performed using the conditioning technique, while relative bias (RBIAS) and relative root mean square error (RRMSE) were used to measure the performance of the estimators. Simulation results and application to real-life data shows that the Zelterman-DLD consistently outperforms the Zelterman-POIS, exhibiting lower RBIAS and RRMSE across all scenarios.

Keywords: Zelterman-DLD, Zelterman-POIS, One-inflated Poisson distribution, Variance estimation

EVALUATING LOGISTIC REGRESSION AND RANDOM FOREST MODELS FOR PREDICTING DIABETES

A. Adesina Oluwaseun — 2025-04-09

Diabetes poses a significant global health challenge, necessitating effective predictive models for early diagnosis and intervention. This study evaluates the performance of logistic regression and random forest models using a dataset comprising information of 390 respondents which was extracted from the data. world to predict diabetes based on health biomarkers such as cholesterol levels, glucose concentrations, BMI, and blood pressure. Results indicate high performance for both models, with the logistic regression model achieving an Accuracy of 91%, Precision of 94%, Sensitivity of 95%, and Specificity of 75%. The random forest model yielded an Accuracy of 89%, Sensitivity of 92%, Precision of 93%, and a Similar Specificity of 75%. The logistic regression model outperforms the random forest in Accuracy, Precision, and Sensitivity, showing greater efficacy in distinguishing between diabetic and non-diabetic individuals.

Keywords: Diabetes, predictive models, logistic regression, Accuracy, random forest

OCCUPATIONAL INCIDENTS AND THE ROLE OF DATASHEETS IN WORKFORCE SAFETY AND RISK MANAGEMENT IN DEVELOPING NATIONS: A THEMATIC REVIEW

Patrick Azodo Adinife — 2025-04-09

Underreporting of workplace incidents is still a major problem in many developing countries, especially in blue-collar industries where safety and risk management deficiencies are still present. By raising safety standards, strengthening risk mitigation techniques, and promoting openness, occupational datasheets provide an organized method of resolving these shortcomings. In order to monitor workplace dangers, evaluate employee capacities, and direct focused actions, this review examines the function of standardized occupational datasheets. It finds important obstacles and chances to use data-driven insights to enhance safety procedures, rehabilitation techniques, and policy creation by examining the body of literature already available on occupational health and safety standards. The findings reveal that while datasheets significantly improve hazard identification and compliance, their adoption is hindered by limited digital infrastructure, lack of training, and resistance to change in traditional industries. Case studies highlight that firms utilizing structured safety documentation experience a measurable decline in workplace accidents and enhanced emergency response preparedness. To bridge the gap between developing countries and international workforce safety norms, the assessment also emphasises how important it is to connect incident reporting systems with best practices from throughout the world.

Keywords: Underreporting, datasheets, safety, risk management, developing nations

CURE FRACTION MODELS BASED ON A LOMAX-EXPONENTIAL DISTRIBUTION WITH APPLICATIONS.

Godfrey Ieren Terna — 2025-04-09

Research has proven that due to the development of new drugs, some patients in a cohort of cancer patients are cured permanently, and some are not cured. The patients who are cured permanently are called cured or long-term survivors while patients who experience the recurrence of the disease are termed as susceptible or uncured. Cure fraction models are usually used to model lifetime time data with long-term survivors. This paper presents a maximum likelihood estimation and analysis of a three-parameter Lomax-exponential distribution (LED) involving a cure fraction parameter with application to censored dataset. In order to capture the proportion of cured patients, a mixture and a non-mixture cure models formulation methods are employed. To assess the usefulness of these models in real life applications, the paper used a real-life dataset on acute lymphoblastic leukaemia (ALL) data. The results revealed that the estimates of the cured proportion based on LED are higher for treatment group I than group II which implies a higher probability survival for patients receiving treatment I than those receiving treatment II. It is also revealed that the estimates of the cured proportion are higher for the mixture cure model than the non-mixture cure model. Furthermore, the study revealed that the mixture cure model based on LED has lower values of AIC and BIC than the non-mixture cure model and LED, meaning that the mixture cure model fits the data better than the non-mixture cure model.

Keywords: Cure fraction, Cure model, mixture, non-mixture, LED, Estimation and application.

A BURR X-PERKS DISTRIBUTION: PROPERTIES AND APPLICATIONS

Godfrey Ieren Terna — 2025-04-09

Probability distributions and associated properties have been used extensively over the years for modelling real life problems, however, most conventional distributions do not adequately analyze many of the skewed real-life datasets and therefore the need for compound or extended probability models. This paper presents a study on a new extension of the Perks distribution by adding one shape parameter to the conventional Perks distribution using the Burr X-G family of distributions. This study has derived and investigated some statistical properties of the Burr X-Perks distribution such as moments, moment generating function, the characteristics function, quantile function, survival function and hazard function. Some plots of the distribution and the reliability functions were generated and interpreted appropriately. The results from the curves show that the distribution is skewed with many shapes depending on the values of the parameters. The plot of the survival and hazard functions shows that the distribution can be used to model time-dependent events, where probability of survival decreases with time, while that of failure increases with time. The parameters of the new model have been estimated using the method of maximum likelihood estimation. The paper evaluated the performance of the proposed Burr X-Perks distribution using two real life datasets and the results revealed that the proposed Burr X-Perks distribution fits the two real life datasets better than the three other distributions considered in this study.

Keywords: Burr X-G family, Perks distribution, Properties, Estimation and application.

POPULATION CHARACTERISTICS IMPACT ON THE WELL-BEING OF PREGNANT WOMEN AND MOTHERS: INSIGHTS FROM THE 2018 NIGERIA DEMOGRAPHIC AND HEALTH SURVEY

O. Amoo Emmanuel — 2025-04-09

The study assessed maternal health, a critical public health issue that serves as one of the challenges in Nigeria's health sector with Nigeria having the highest maternal mortality and morbidity at the global comparison. Maternal health refers to women's " health during pregnancy, childbirth, and postpartum”. Maternal Mortality and Morbidity Rates in Nigeria are the highest among the countries globally (Azuh et al., 2017). The study examined the Population Characteristics shaping maternal health outcomes, exploring data from the 2018 Nigeria Demographic and Health Survey (NDHS). The study systematically evaluates vital population characteristics variables such as Educational Attainment, Socioeconomic Status/Wealth Index, and Geographical Location, which significantly influence access to healthcare services, antenatal care (ANC) utilization, and the overall incidence of Maternal Mortality (MM). Through a laborious analysis of these factors, the study highlights the disparities in maternal health outcomes across different demographic groups and regions in Nigeria. The study offered policy recommendations, such as strengthening maternal health education and awareness, collaborating with NGOs and local leaders to conduct workshops on the importance of antenatal care (ANC), birth preparedness, and recognizing danger signs during pregnancy, and enhancing economic support for vulnerable women. Expanding social protection programs, such as conditional cash transfers, can effectively subsidize maternal healthcare costs for low-income women, as noted by the World Health Organization and World Bank Group (2023). In addition, strengthening Nigeria’s National Health Insurance Scheme (NHIS) to cover ANC, delivery, and emergency obstetric care for rural populations is recommended (Amedari & Ejidike, 2021). Improving healthcare infrastructure in rural areas is another vital component of these recommendations. Increasing funding for primary healthcare centers (PHCs) in rural regions will ensure the availability of skilled birth attendants and essential medications (Oluwole et al., 2022). Increasing access to healthcare services in rural areas and improving the quality of ANC, based on the empirical evidence to address the identified gaps, improve maternal health services, and reduce preventable maternal deaths nationwide.

Keywords: Demographic determinants, Maternal health, Nigeria, NDHS, Antenatal care, Maternal mortality

ON BIVARIATE EXPONENTIAL DISTRIBUTION BASED ON ALI-MIKAILHAQ COPULA FUNCTION

Aliyu Yakubu — 2025-04-09

Copula function has become one of the most popular methods in constructing bivariate distributions. In this article a new bivariate generalized exponential distribution based on Ali-MikailHaqs copula function is introduced. Estimation of the parameters of the Ali-MikailHaq bivariate generalized exponential distribution was obtained via Bayesian method of estimation. An application of the proposed methodology was illustrated by fitting the distribution to a survival data set and compares its performance with other competing distribution. Based on the deviance information criteria (DIC) values, it is shown that, the Bivariate generalized exponential Ali-Mikha’il-Haq distribution is more efficient.

ON BURR X NON-MIXTURE LONG-TERM SURVIVAL MODEL WITH APPLICATIONS TO MEDICAL DATA

Aliyu Yakubu — 2025-04-09

In the analysis of lifetime data, it is usually assumed that each and every subject in the study population will experience the event of interest if followed for a long period of time. However, due to improvement in the field of medicine, some subjects stay longer than expected. These subjects are termed long-term survivors. Hence, this study introduces a long-term survivor model based on Burr X distribution. The parameters of the model were estimated the Bayesian estimation procedure. To illustrate the applicability of the introduced methodology, a medical dataset was used fitted and the result compared with that of the model without the proportion of long-term survivors.

Keywords: Survival model, Long-term survivor model, non-mixture long-term survivor model, Burr X distribution.

COMPARATIVE ANALYSIS OF HAAR AND DAUBECHIES STATISTICAL WAVELETS FOR FINANCIAL TIME SERIES ANALYSIS

W. B. Akinkunmi — 2025-04-09

This study presents a comparative analysis of Haar and Daubechies statistical wavelets for analyzing financial time series data. We examine the application of these wavelets in extracting insights from financial data, including trend, fluctuations, energy distribution, and complexity. Our results show that both Haar and Daubechies wavelets can effectively capture the characteristics of financial data, but they differ in their methodological approaches and estimates. The Haar wavelet analysis reveals a gradual increase in the trend of the S&P GREEN BND SELECT INDEX - PRICE INDEX data, with a moderate level of complexity or uncertainty. The energy distribution of the Haar wavelet coefficients shows that the majority of the energy is concentrated in the low-frequency components. Our analysis demonstrates the effectiveness of statistical wavelets in extracting insights from financial data, which can be used to inform investment decisions, risk management strategies, and other financial applications.

Keywords: Statistical Wavelets, Haar Wavelet, Daubechies Wavelet, Financial Time Series Analysis, Signal Processing.

COMPARATIVE ANALYSIS OF MinRV AND MedRV MEASURES FOR VOLATILITY ESTIMATION AND JUMP DETECTION

W. B Akinkunmi — 2025-04-09

This study presents a comparative analysis of two volatility measures, Minimum Realized Volatility (MinRV) and Median Realized Volatility (MedRV), in assessing financial market dynamics. We examine the application of these measures in detecting significant price movements and volatility using data from the Nigerian Stock Exchange. Our results show that both MinRV and MedRV measures can be effective in capturing volatility, but they differ in their methodological approaches and estimates. The MinRV measure yields an estimated daily volatility of 2.0% and an annualized MinRV of 28.3%, while the MedRV measure provides stock-specific volatility estimates, with ABCTRANS showing 4.67% volatility. Our analysis highlights the importance of considering multiple volatility measures and methodologies in financial analysis. The findings recommend that investors, policymakers, and other stakeholders consider using a combination of MinRV and MedRV volatility measures to gain a more nuanced understanding of financial market dynamics and identify potential risks and opportunities.

Keywords: Volatility Measures, MinRV, MedRV, Financial Markets, Risk Management, Jump Detection.