DEPARTMENT OF STATISTICS

This study investigates the relationship between crop performance, fertilizer application, and soil types using multivariate statistical analysis. The main objective is to determine how different fertilizer types and the rate of application, in combination with soil characteristics, influence major growth and yield parameters of crops. Data were collected on soil properties
(such as pH, organic matter, nitrogen, phosphorus, potassium, and texture) and crop growth parameters (including germination percentage, plant height, number of leaves, leaf area, biomass, and yield). The canonical correlation analysis was employed to identify patterns and quantify the strength of associations among these variables. The results revealed that soil fertility factors and fertilizer applications significantly influenced crop growth and yield performance, with organic matter and fertilizer rate producing optimal responses. The analysis demonstrates the usefulness of Canonical Correlation Analysis in handling complex agricultural data.

Nosocomial infections, also known as hospital-acquired infections, are bacterial infections contracted within healthcare settings. They contribute significantly to the burden of disease by prolonging hospital stays, increasing treatment costs, complicating surgical outcomes, and, in severe cases, leading to death. Methicillin resistant staphylococcus aureus (MRSA) is the most isolated
pathogen of nosocomial infections and the most studied in the literature. Despite their impact, awareness of nosocomial infections remains limited, leaving patients, healthcare workers, visitors, and even family caregivers vulnerable to its transmission. In low-income and middle-income countries the practice of a family caregiver assisting an inpatient is common, and they can contact and transmit infections while carrying out various activities within the hospital environment. Deterministic and stochastic models have been widely applied to understand the transmission dynamics of nosocomial infections and provide valuable insights. However, existing models have often overlooked the role of patients’ family caregivers, who can act as important but underrecognized vectors of transmission. In this thesis, deterministic and stochastic models that explicitly incorporate family caregivers as a distinct transmission pathway of methicillin-resistant Staphylococcus aureus (MRSA) are developed. For the deterministic framework, the basic reproduction number �0 is derived along with the conditions for disease-free and endemic equilibria. The stochastic framework developed using a Continuous-time Markov Chain (CTMC) extends the deterministic model by incorporating random fluctuations through its drift and diffusion terms. This provides deeper insight into the system’s variability, extinction probabilities, and outbreak risks that cannot be fully captured by the deterministic model. For the deterministic model, the basic reproduction number is evaluated and subjected to sensitivity analysis, using plausible parameter values from surveillance studies within Nigerian hospitals, whereby revealing the dominant influence of hand-hygiene compliance of caregivers and healthcare workers, as well as decontamination rates of both caregivers and healthcare workers. The stochastic simulation in MATLAB gives the stochastic sample paths, time-series behaviours of the state variables and extinction probability. The numerical results illustrate that while the deterministic model captures mean epidemic behaviours, stochastic models reveal substantial variability and probability of infection extinction especially in settings with effective hand hygiene compliance of caregivers and healthcare workers. These analyses reveal the importance of integrating patients’ family caregivers in modeling the spread of MRSA in hospitals.

The exponential distribution became the cornerstone of survival analysis and reliability engineering throughout the latter half of the 20th century and it is imperative to mention that the Successive times in exponential distribution are assumed to occur independently and randomly over time with a constant rate. The aim of the study was to develop a generalized and flexible bivariate exponential distribution that will incorporate correlation parameter 𝜌, extending the domain to a positive real-line using the framework of linear regression. A secondary dataset from Federal Road Safety Corps on road accidents in Imo State from 2020 to
2024 was used in the study and it was obtained from the Head office of Federal Road Safety Corps, off Egbu road Owerri, Imo state. In the study, we developed a generalized and bivariate exponential model that incorporates a correlation parameter, while preserving analytical simplicity. The proposed model, referred to as the New Correlated Bivariate Exponential Distribution (NCBED). The consistency of the NCBED was assessed using Kolmogorov Smirnov and Cramer Von Mises tests, in comparison with the baseline Grine model (2018). The Federal Road Safety (FRSC) dataset demonstrates that both injury and fatality data follow heavy-tailed exponential-type distributions and the NCBED provided a superior fit compared to the baseline model, capturing real-world correlations between crash outcomes. The findings indicate that the Maximum Likelihood estimates of the proposed model are consistent with the nature of the model.

The research work explores the comparison of various methods for estimating variance components in a two-way random effects model, a critical task in experimental data analysis. The methods assessed include classical Analysis of Variance (ANOVA), Restricted Maximum Likelihood (REML), and Bayesian estimation. The experiment was designed with treatments (3 levels) and blocks (4 levels), with each combination replicated 5 times, resulting in 60 observations. The objective was to estimate variance components attributable to treatments, blocks, and errors. The results were compared across the three methods: ANOVA produced variance components of σ²α = 3.84, σ²β = 2.43, and σ²ε = 3.58, while REML and Bayesian estimates were σ²α = 4.805 and 4.75, σ²β = 2.4067 and 2.60, and σ²ε = 3.58 and 3.60, respectively. While the three methods yielded similar results, minor differences were observed, reflecting their respective properties. ANOVA, though simple and interpretable, may be biased in small samples or unbalanced designs, whereas REML offers better performance in such situations, and Bayesian estimation provides flexibility with credible intervals to quantify uncertainty. The research work highlights the importance of method selection depending on sample size, design, and the need for uncertainty quantification, suggesting future work on more complex or larger-scale experiments.

The work is intended to study the moments of distributions and in particular pareto-II distribution named after Italian scientist, Vilfred pareto. This study was guided by the following objectives; to obtain the rth moment, obtain the mean, variance, skewness and kurtosis of the pareto-II distribution and finally obtain some numerical results of the moment. The study employed the knowledge of differential and integral calculus with transformation of variables to obtain several expressions as we shall be seeing. Statistical software "R" was used to run analysis and obtain numerical results of the moments. Finding revealed that the parameters of the distribution are important in determining the behavior of the moments. From the findings, it implied that the study of moments is important and applicable to study of distributions. Keywords Moments, Parameter, Distribution, Mean, variance, skewness, kurtosis.

Understanding customer’s preference for product features is crucial for businesses seeking to optimize their offerings and increase sales. This study aims to identify key product features that significantly influence purchasing decisions. The researchspecifically examines customer preference across different demographic groups, such as age and income levels, to determine variations in product feature importance. Using chi-square test for analysis, this study evaluates the association between demgraphic factors and product features preferences, providing insight into customer’s decision-making patterns. Based on the analysis it was found that the key product features that affect customers purchasing decision is quality and performance. The findings will help businesses tailor their marketing and product development strategies to better align with costumer expectations.

This research undertakes a comprehensive statistical analysis of Nigeria's Gross Domestic Product (GDP) spanning a decade, with a focus on estimating Autoregressive (AR) models using two prominent statistical methods: the Yule-Walker method and the Least Squares method. The study aims to provide statistical insights into the underlying dynamics of Nigeria's economic performance during this period. The research commences by delineating the statistical framework of ARmodels, which offer a statistical representation of a time series based on its past values. Subsequently, the Yule-Walker method is introduced, a statistical technique leveraging autocorrelation functions to estimate AR model parameters. The statistical properties of Yule-Walker estimators are elucidated in the context of Nigeria's GDP data. In contrast, the Least Squares method is presented as an alternative statistical approach, characterized by its objective to minimize the sum of squared prediction errors. A statistical framework for the least squares estimators is outlined, providing insights into the statistical properties of parameter estimates and their significance in explaining variations in Nigeria's GDP. The core of the research involves the statistical analysis of Nigeria's GDP time series data over the 10-year period. Both the Yule-Walker and Least Squares methods are applied to estimate AR models tailored to the GDP data. The statistical comparison is based on goodness-of-fit statistics, such as the Akaike Information Criterion (AIC), to evaluate the models' adequacy in capturing the statistical patterns within the GDP dataset.

his study looks into the spending behavior of students at the University of Benin, Nigeria, in order to establish the relationship between demographic factors such as gender, age, and academic level on expenditure behaviour. The purpose of the study was to gain a better understanding of how students’ money is spent in areas of food, transport, housing, academic materials, and personal expenses. A quantitative approach was used in the study where 228 questionnaires were completed and descriptive and inferential analysis were employed on the data collected. The results show that the amount of money spent by the students on food, transport, and housing is highest. The results of the statistical analysis indicate that there is no significant association between the gender of the respondent and the primary source of income, however, more males are likely to save than females. The result of the study shows that the academic level affects the spending and higher-level students spend more money on the necessities. The transport expense also depends on the age of students, the older students spend more money on transport.
The results conclude that, among students, academic level and age are important predictors of spending behaviour; the impact of gender is limited to saving behaviour. For practice, it is suggested that financial literacy be enhanced, housing and transportation be made cheaper, and financial aid be increased. Further research could be directed towards the patterns of spending at different universities, changes over time, the effect of financial aid and part-time employment on spending. This research can help to explain student financial problems and offers some practical suggestions for students, administrators and policymakers.

This study examines the influence of Artificial intelligence on jobs: unemployment and displacement. It captures the implication of Artificial intelligence on the workforce as a whole. The Literature review of the study was segmented into three(3) sections namely; Conceptual review, Theoretical review, and Empirical review. Major Statistical tools of analysis used includes; data visualizations using histograms and the multinomial logistic regression. All tests done were conducted at the 0.05 level of significance. Major findings show that python, AI algorithms, certifications, problem solving, mathematics, and the other skills have a significant effect on getting an AI-related job. The study concludes that the impact AI has on jobs, is more on job loss as regards low-skilled workers, however AI has the ability to complement human workers, which will in turn lead o increased efficiency and productivity. Also as regards job displacement, it will affect mainly low-skilled and routine jobs.