http://ir.tum.ac.ke/handle/123456789/17344 Contains electronic theses & dissertations for this department Sat, 13 Jun 2026 18:39:31 GMT 2026-06-13T18:39:31Z http://ir.tum.ac.ke/handle/123456789/17673 A New Parametric Yang-Prentice Regression Model With Applications to Real-Life Survival Medical Data With Crossing Survival Curves ISHAG, MOHAMED A. S.; ELBATAL, IBRAHIM; WANJOYA, ANTHONY KIBIRA; ADEM, AGGREY; ALMETWALLY, EHAB M.; AFIFY, AHMED Z. The Yang and Prentice (YP) regression models have attracted considerable attention in the scientific community due to their ability to handle survival data with crossing hazard functions. These models encompass both the proportional hazards (PH) and proportional odds (PO) models as special cases. A key feature of the YP framework is the inclusion of distinct short-term and long-term hazard ratio parameters, which allow it to accommodate intersecting survival curves. Notably, the original YP model leaves the baseline hazard function unspecified. In this study, a fully parametric method is introduced for fitting the YP model within a general regression context. The core idea involves modeling the baseline hazard using the exponentiated-Weibull distribution, which provides both the flexibility of parametric modeling and analytical tractability. To assess the effectiveness of the proposed approach, comprehensive simulation studies were performed. The results indicate that the model performs robustly even with moderate sample sizes and demonstrates improved accuracy compared to the original YP model, particularly in general regression scenarios beyond the traditional two-sample setup. Additionally, the utility and effectiveness of the proposed method are illustrated through applications to real-world datasets. The results underscore the model’s strengths in capturing complex survival patterns and enhancing the analysis of survival data. DOI: 10.1109/ACCESS.2025.3601730 Fri, 22 Aug 2025 00:00:00 GMT http://ir.tum.ac.ke/handle/123456789/17673 2025-08-22T00:00:00Z http://ir.tum.ac.ke/handle/123456789/17672 Volatile constituents of different smoking materials in bee honey harvesting and detection of mean differences in calming time using completely randomized design Korir, Thomas Kiptanui; Adem, Aggrey Onago; Odalo, Josiah Ochieng Beekeeping and honey production provide a wide range of economic contributions through income generation from marketing of honey and related products, and creation of non-gen- der-biased employment opportunities. Traditionally, honey harvesting has been accom- plished by making use of naked flames to rid of or even destroy honey bees or employing smoke to suppress their aggression. However, there is limited data on the effectiveness of different smoking materials used to calm honey bees during harvesting, necessitating research into establishing effective options that maximize honey yield while preserving the bee colony. This study evaluated the adaptability of variation in time taken to calm the honey bees, Apis mellifera scutellata, during honey harvesting using two fungi species Calvatia cyathiformis and Calvatia gigantean, air-dried hay from Chloris gayana and Nasiwa setaria, and the bark of Juniperus procera (African pencil cedar) as smoking agent sources. Gas chromatography-mass spectrometry (GC-MS) analysis of the volatile compounds released by the smoldering smoking materials identified naphthalene, 2-(2 octenyl)-cyclopentanone, 1,3-diethylbenzene, 2-methoxyphenol, 1,5,5-trimethyl-6-(3-methyl-buta-1dienyl)-cyclohexene, creosol, 4-ethyl-2-methoxyphenol, 2,4-bis(1,1-dimethylethyl)-phenol and cedrol as the major components of the calming smokes. Completely randomized design was used to compare the mean calming time among the smoking materials. Significant differences in mean calm- ing time among the groups was reported. Post -ANOVA analysis using Turkey’s HSD test revealed significant differences among the smoking material types, with C. cyathiformis and C. gigantean emerging as the most effective calming agents. DOI: https://doi.org/10.1080/00218839.2025.2531310 Fri, 01 Aug 2025 00:00:00 GMT http://ir.tum.ac.ke/handle/123456789/17672 2025-08-01T00:00:00Z http://ir.tum.ac.ke/handle/123456789/17671 Predicting Student Attrition in Kenyan Universities: A Comparative Analysis of Machine Learning Algorithms Nyawira, Lilian; Musau, Obadiah; Adem, Aggrey; Jobunga, Eric One of the primary goals of higher education institutions is to provide high-quality education and ensure a high completion rate. Reducing student attrition is one strategy for attaining high-quality education. Identifying students who are susceptible to dropping out and the variables that lead to dropouts are essential to achieving this. The purpose of this research was to ascertain how machine learning models might be used to forecast student attrition in Kenyan universities. Based on a number of classification criteria, such as F1 score, precision and accuracy, the study assessed and contrasted the performance of numerous algorithms, including Decision Trees, Random Forest, Naive Bayes, and Logistic Regression. The analysis demonstrated how well Logistic Regression worked, outperforming the other models and consistently striking a balance between precision and recall. Decision Trees and Random Forest, despite showing improvements through hyperparameter tuning, still struggled to identify students at risk of attrition. Naive Bayes, while relatively balanced, did not match the performance of Logistic Regression. The study provided a comprehensive overview of each model's strengths and limitations and suggests future work to further optimize the models for better predictive performance. DOI: https://doi.org/10.62049/jkncu.v5i2.313 Wed, 01 Jan 2025 00:00:00 GMT http://ir.tum.ac.ke/handle/123456789/17671 2025-01-01T00:00:00Z