AAU Institutional Repository (AAU-ETD)
Addis Ababa University Institutional repository is an open access repository that collects,preserves, and disseminates scholarly outputs of the university. AAU-ETD archives' collection of master's theses, doctoral dissertations and preprints showcase the wide range of academic research undertaken by AAU students over the course of the University's long history.
How to Submit Your Work
The repository contains scholarly work, both unpublished and published, by current or former AAU faculty, staff, and students, including Works by AAU students as part of their masters, doctoral, or post-doctoral research
- All AAU faculty, staff, and students are invited to submit their work to the repository. Please contact the library at your college.
You may contact digirep@aau.edu.et.with any questions about the repository
Colleges,Institutes in AAU-ETD
Select a college,institute to browse its collections.
Recent Submissions
Data Driven Quality Assurance in Agile Software Project: Optimizing Scrum Checkpoints through a Conceptual Framework
(Addis Ababa University, 2025) Nigusu Daniel; Ayalew Belay
This thesis investigates the enhancement of quality assurance (QA) in agile software development, with a specific focus on the Scrum framework. Agile methodologies, while effective in promotingflexibility and fast delivery, often face challenges in maintaining consistent software quality due to limited time for testing and the absence of structured QA practices. Traditional quality modelsare often incompatible with the iterative and dynamic nature of agile processes.To address this issue, the study proposes a data-driven conceptual framework that integrates quality assurance checkpoints into Scrum-based agile workflows. The framework is built upon twoestablished quality models: the Plan–Do–Check–Act (PDCA) cycle and the Cross-IndustryStandard Process for Data Mining (CRISP-DM). These models guide the systematic placementand evaluation of QA checkpoints across the Software Development Life Cycle and the ProjectLife Cycle.The research followed a mixed-methods approach. The framework was initially developed throughan in-depth literature review and expert consultations (qualitative phase), and then empiricallyvalidated using a quantitative survey of 51 agile professionals in Addis Ababa. The structuredquestionnaire captured data on QA practices across Scrum phases, and statistical analysis usingSPSS including correlation analysis was applied to identify critical checkpoints and evaluate theframework’s relevance.Results show that integrating structured, data-driven QA checkpoints into Scrum significantlyenhances defect tracking, test planning, and product quality. The proposed framework provides ascalable and practical solution for organizations seeking to align agile delivery with rigorousquality standards.
The Impact of Red Billed Quela (Quela quela) on Agricultural Products in Merti Woreda, Oromia Region, Ethiopia
(Addis Ababa University, 2025-08-01) Mustefa Abadir; Bezawork Afework
In Ethiopia crop production is the leading sub-sector of agricultural economy where more than 60% of agricultural GDP originated from crop production. The objective of this research was to assess the impact of Red Billed Quela (Quela quela) on Agricultural Products in Kola Kebele of Merti woreda, Oromia region, Ethiopia. Quantitative research methodology was employed to collect data where structured questioner was used to 100 farmer household respondent selected using simple random sampling techniques. The data was analyzed by tables percentage and frequency. The results showed that quela bird infestation is a major barrier to food security and livelihood sustainability in Merti woreda, and that improved, integrated, and institutionalized control strategies are needed. Mixed farming systems, productivity levels remain moderate, with most farmers harvesting only 11–15 quintals per year, reflecting both the small landholding size and recurrent birdrelated losses. Farmers largely depend on traditional control methods such as scaring and with 70% applying all available methods. Institutional and government measures were found to be weak or inconsistent, forcing communities to rely heavily on indigenous knowledge. In addition, limited irrigation infrastructure, poor land accessibility near water sources, and inadequate extension services further constrain productivity improvements. Only 4% of respondents had experience with irrigated sorghum cultivation, and 30% had received guidance from agricultural institutions. Both season of bird outbreak and harvesting time in autumn causes vulnerability and making farmers highly exposed to crop loss. Farmers recomanded that strengthening government intervention,facilitating the society at the time of bird migration to prevent the severity of the cereal crops additionally enhance the traditional control methods for best management of the quela birds.
Evaluation of the Impacts of Land Use/Land Cover and Climate Change on Flood Occurrence in the Gumara Watershed Upper Blue Nile Basin Ethiopia
(Addis Ababa University, 2025-08-01) Haile Belay; Assefa Melesse; Getachew Tegegne
Floods are among the most destructive natural hazards worldwide, with their frequency and intensity exacerbated by changes in land use/land cover (LULC) and climate. Vegetation cover and climatic variables are also interlinked, further influencing hydrological processes. Thus, this study evaluated the impacts of LULC and climate changes on flood occurrence and explored the relationships between vegetation cover and climate variables. Extreme rainfall analysis and vegetation‒climate relationships were assessed at the basin scale in the Upper Blue Nile (UBN) Basin of Ethiopia, whereas LULC change detection and flood modelling were conducted at the watershed scale in the Gumara watershed, which is located in the UBN Basin. The study comprises four interrelated specific objectives. First, the performances of eight CMIP5 and eight CMIP6 general circulation models (GCMs) in reproducing observed extreme rainfall indices (ERFIs) for the baseline period (1981–2005) were evaluated. The GCMs were assessed via statistical metrics (correlation, RMSE, and PBIAS) and ranked via the technique for order preferences by similarity to ideal solution (TOPSIS). The four best-performing models (MIROC5, MRI-CGCM3, CanESM2, and MPI-ESM-LR from CMIP5 and MPI-ESM1-2-LR, MRI-ESM2-0, CNRM-CM6-1, and BCC-CSM2-MR from CMIP6) were bias-corrected, and their ensembles were used for projecting extremes for the future period (2031–2080) under the RCP4.5, RCP8.5, SSP2-4.5, and SSP5-8.5 scenarios. Second, historical and future LULC changes were analysed in the Gumara watershed. LULC maps for 1985, 2000, 2010, and 2019 were generated via a random forest classifier on the Google Earth Engine (GEE) platform, which achieved high overall accuracies. Future LULC maps for 2035 and 2065 under business-as-usual (BAU) and governance (GOV) scenarios were projected via the Cellular Automata–Markov (CA-Markov) model. The results revealed considerable expansion of cultivated land and settlements under BAU and a relative increase in forest cover under GOV. Third, rainfall-runoff modelling was conducted in the Gumara watershed via the Hydrologic Engineering Center-Hydrologic Modelling System (HEC-HMS) for various return periods and combined LULC and climatic scenarios. CombinedThe combined impacts of LULC and climate changes on annual maximum (AM) flows were quantified, indicating an increasing trend under future scenarios. In particular, the highest flow is expected under the SSP5–8.5 (2056–2080) climate scenario combined with the BAU-2065 land use scenario. Additionally, flood source area identification (FSAI) was performed in this part of the study on the basis of the unit flood response (UFR) approach to prioritize subwatersheds contributing most to peak flows. Finally, flood inundation mapping was conducted by integrating synthetic aperture radar (SAR)-based change detection and thresholding methods on the GEE platform and HEC-RAS 2D hydrodynamic modelling. Sentinel-1 SAR images were used to map flood inundation from 2017–2021, and the HEC-RAS model was calibrated and validated via SAR-derived flood inundation maps. Future inundation maps indicated increased flood extents, primarily attributed to LULC and climate changes. Overall, this study provides a comprehensive assessment of how coupled LULC and climate changes influence flood risk and how climatic variables affect vegetation cover. The findings offer a scientific basis for flood risk management, watershed conservation planning, and policy formulation under future climate and land development scenarios. Moreover, the methods and results presented herein provide a valuable framework for similar studies in other regions undergoing rapid environmental and climatic changes
Predicting Risk Factors for Customer Churn with Explainable Artificial Intelligence
(Addis Ababa University, 2025-09-01) Ermiyas Amare; Michael Melese
Customer retention remains a critical challenge in the banking sector, particularly in emerging economies where customer behavior is rapidly evolving. Despite access to extensive data, Bunna Bank still relies on conventional customer relationship man agement practices that are often manual, reactive, and lacking analytical depth. This study focuses on predicting the risk factors for customer churn at Bunna Bank using machine learning and explainable artificial intelligence methods, using a real-world dataset of 308,293 customer records with 19 attributes having transactional behavior, account activity, and demographic information. Six machine learning classifiers were evaluated: Random Forest, Gradient Boosting, XGBoost, LightGBM, Extra
Trees, and Multi-Layer Perceptron. Before modeling, the dataset was balanced using SMOTE to address class imbalance, and a hybrid feature selection strategy was applied to identify the most relevant attributes. XGBoost achieved the highest accuracy (92.23%), along with strong precision (0.96), recall (0.94), F1-score (0.95), and AUC (0.97), making it the most effective model for identifying high-risk customers. To enhance interpretability, SHAP and LIME were applied to the XGBoost model, providing both global and local explanations. The analysis revealed that features such as Has Mobile Banking, Is USSD Active, Transaction Count, Transaction Recency, Age, and ATM Activity were the most influential drivers of churn risk. Active digital banking usage and recent transactions were protective factors, while younger customers with low engagement showed higher churn risk. A prototype was de
veloped using Streamlit to risk factor prediction that integrate predictive modeling with XAI enables accurate and transparent into customer churn
Comparison of the Efficiency of CSF and Serum PCR in the Diagnosis of Bacterial Meningitis Among Suspected Cases at Tikur Anbessa Specialized Hospital and Yekatit 12 Hospital Medical College, Addis Ababa Ethiopia
(Addis Ababa University, 2025-02-27) Lema Ayele; Daniel Asrat; Getachew Tesfaye
Background
Early and accurate diagnosis of bacterial meningitis reduces mortality and sequelae. In resource-limited settings like Ethiopia, where laboratory diagnostics are scarce, developing a diagnostic tool based on an easy-to-collect blood sample for the diagnosis of bacterial meningitis is crucial to improving the diagnosis and prompt treatment initiation that has a great impact on patient outcomes.
Objective
To compare the efficiency of cerebrospinal fluid and blood PCR for the diagnosis of bacterial meningitis in patients suspected of meningitis.
Methods
An institutional-based cross-sectional study was conducted at Tikur Anbessa Specialized Hospital and Yekatit 12 Hospital Medical College from August 2023 to February 2024. A total of 202 patients who were suspected of having meningitis were given their consent to collect the demographic, clinical data, CSF, and blood samples. Demographic data and clinical samples (CSF and Blood) were collected by using pre-tested standardized questionnaires and aseptic techniques, respectively. Conventional laboratory analysis and bacterial isolation using aliquots of CSF were performed following the standard microbiology laboratory procedures. The leftover CSF and serum samples were stored at -20°C until transported to the Armauer Hansen Research Inistitute (AHRI) for further molecular tests. Genomic DNA from CSF and serum was extracted by the Bioer NPA-32P automated nucleic acid purification machine. A multiplex PCR technique was used to detect the presence of Neisseria meningitidis, Haemophilus influenzae, Streptococcus pneumoniae, Streptococcus agalactiae, Escherichia coli, Staphylococcus aureus, Listeria monocytogenes, and Klebsiella pneumoniae. The end product of the PCR amplicon was observed under the Vilber machine after staining with ethidium bromide. Data were entered into Epi Info 7, cleaned, exported to Excel, and loaded into SPSS version 26 software for analysis. Descriptive statistics were presented in tables and bar charts. The agreement between the tests to detect bacterial meningitis was calculated by using an SPSS contingency table.
Result
Out of 202 study participants, 54% (n = 109) were males, and 88.6% (n = 179) of them were from the pediatric population, with an average age of 1.45 ± 0.69 years. Reduced ability to breastfeed was the most common clinical presentation 74.8% (n = 151), followed by high fever 71.3% (n = 144) and a loss of consciousness 66.3% (n = 134). Of 202 CSF, 5.4% (n = 11) and 4.5% (n = 9) of the samples tested were positive for gram stain and culture, respectively. Overall, bacterial DNA was detected in 26% (n = 54) and 46% (n = 93) CSF and serum samples, respectively. The concordance between CSF and serum PCR was observed in 34 (17%) cases. E. coli was the most predominant bacteria detected in both CSF and serum PCR at 55.5% (n = 30) and 61.2% (n = 57), respectively, followed by S. pneumoniae detected in both CSF and serum PCR at 15% (n = 8) and 29% (n = 27), respectively. The overall agreement between CSF and serum PCR is 60.9% with a positive percent agreement, and negative percent agreement between the two tests is 62.9% and 60.1%, respectively.
Conclusion
Bacterial DNA was 20% more frequently detected in serum compared to CSF, and 42% compared to culture in suspected cases of meningitis. Furthermore, these findings are clinically very important to minimise the invasive lumbar puncture for CSF collection. Overall, serum PCR could be considered as a supplementary diagnostic tool, a less invasive diagnostic procedure for the diagnosis of bacterial meningitis. However, further similar studies are important using qPCR and both blood and CSF culture.