http://ir.haramaya.edu.et//hru/handle/123456789/10 Fri, 22 May 2026 23:15:42 GMT 2026-05-22T23:15:42Z http://ir.haramaya.edu.et//hru/handle/123456789/8419 MACHINE LEARNING-BASED PREDICTION OF UNDER-FIVEMORTALITY USING HEALTH,SOCIO-DEMOGRAPHIC,ANDCLIMATE DATA Feyisa Abebe; Abebe Belay Adege (PhD); Mr. Tadesse Kebede Health is a state of full well-being and a cornerstone of international development, withconsiderable investments made over the last three decades to reduce morbidity and mortality. Under-five mortality, which is generally defined as the death of children under the age of five, is still a critical public health challenge in developing countries. The study here proposesamachine learning-based approaches for predicting under-five mortality using health, socio-demographic, and climate data from Eastern Hararghe, Ethiopia. The data used in this study were collected from the Hararghe Health DemographicSurveillance System and Ethiopian National Meteorology Agency. The followingeight supervised machine learning algorithms were considered: Naïve Bayes (NB), Support VectorMachine (SVM), K-Nearest Neighbors (KNN), Decision Tree (DT), RandomForest (RF), eXtreme Gradient Boosting (XGBoost), Attentive Tabular Network (TabNet), andConvolutional Neural Network (CNN). The proposed framework covers data preprocessing, exploratory data analysis, model training, prediction, performance evaluation, andidentification of the key determinants of under-five mortality. It was observed that an80:20split produced an optimum performance in the models. Preprocessing techniques werethenapplied to enhance data quality before training the machine learning models. There weretwoexperimental setups: one with a data-balancing technique and the other without.Results indicated that balanced datasets always outperformed. Amongst all the modelsdeveloped, the XGBoost recorded the highest accuracy, having testing accuracy scoreof97.9%, precision of 98%, F1-score of 98%, and recall of 98%. The determinants of under-fivemortality identified in this study were antenatal care, child gender, wealth index, total numberof alive children, preceding child alive, physical healthy ,birth place and weight of baby. Intheend, the XGBoost algorithm emerged as the best among other models, proving to be themost reliable predictive model for under-five mortality. This study has shown the potential ofmachine learning providing helps in tackling critical public health challenges by leveragingdiverse datasets to enhance decision-making and interventions. Household-level climatedatawere not utilized in this thesis, which would be taken into account by future researchers. 129p. Tue, 01 Apr 2025 00:00:00 GMT http://ir.haramaya.edu.et//hru/handle/123456789/8419 2025-04-01T00:00:00Z http://ir.haramaya.edu.et//hru/handle/123456789/8303 COMPUTING RISK ANALYSIS OF UNDER-FIVE CHILDREN WITH PNEUMONIA: THE CASE OF GENERAL HOSPITALS IN EAST HARARGE ZONE, ETHIOPIA Mohammed Abdella Abrahim; Dr.Kasahun Takele (PhD); Million Wesenu (Asst. Prof) Pneumonia is the one of infectious cause of morbidity and mortality among children under age five years in low and middle-income countries including Ethiopia. The main objective of this study was to analysis survival time to death under-age five children with pneumonia in the presence of computing risk in East Hararge Zone General Hospitals. To meet the study objective the retrospective study design was used secondary data on 436 sampled under-age five children with pneumonia patients from January 1st, 2022 up to January 1st, 2023. Out of 436 patients, 35 (8.03%) died from pneumonia, 29 (6.65%) died from other causes, 292 (66.97%) recovered from pneumonia, and 80 (18.35%) were censored. The maximum relative difference observed for the covariate between the cause-specific hazard ratios and sub-distribution hazard ratios was 89%. The model comparison was done using the sub-bayesian information criteria to select the best model to fit the data. The cause-specific hazard frailty model was appropriate as compared to candidate models to fit the pneumonia patient’s dataset. There was an unobserved heterogeneity due to clustering (Hospitals) in the survival experience of patients in Eastern Hararghe of General Hospitals. The final results of the cause-specific hazard frailty model showed that sex, age group 12-23, age group 24-35, health insurance, season of diagnosis in summer, acute respiratory tract infection, patient referral status, micro-nutrient deficiency, and weight were significant risk factors associated with death due to pneumonia in thepresence of competing risk. The male patients, patients whose age categories were 12-23, 24-35 months, the season of diagnosis summer, and patients with ARTI had a significantly increased risk of death due to pneumonia patients. While, patients who use health insurance, patients referred from other health centers and weight had significantly decreased risk of death due to pneumonia patients. It is recommended that the hospital variation should be taken into account during intervention and awareness creation should give to the local community to increase the health insurance status which reduces the child mortality risk of pneumonia. 84 Sat, 01 Jun 2024 00:00:00 GMT http://ir.haramaya.edu.et//hru/handle/123456789/8303 2024-06-01T00:00:00Z http://ir.haramaya.edu.et//hru/handle/123456789/8273 DETERMINANTS OF TIME TO DEATH OF STROKE PATIENTS: USING SHARED FRAILTY MODELS KUMELA AYANSA TERESSA; Alebachew Abebe (Asst. Prof); Kassahun Takele (PhD) Stroke is a severe medical condition causing brain cell death, causing physical changes, communication problems, cognitive issues, emotional problems, and pain. Stroke, the second largest global death cause, is causing an increasing burden of mortality, morbidity, and disability. The main aim of this study was to analyze the factors influencing the time to death of stroke patients by using shared frailty models. The study was conducted in Harar City, at Harar General Hospital, Jegol Hospital and Hiwot Fana Specialized University Hospital. The study was utilized a retrospective study design and considered 224 sample stroke patients from 1 September 2020 to 1 November 2023. Among the total of 224 stroke patients 51(22.77%) experienced a death and the rest 173(77.23%) were censored. The estimated median time to death for stroke patients was 14 days, highlighting the acute nature of this condition. Through rigorous analysis, the Inverse Gaussian frailty model with the Weibull baseline hazard function emerged as the most suitable statistical model, accurately predicting the time to death of stroke patients and yielding the smallest AIC value. The study identified hypertension, cardiac disease, diabetes mellitus, atrial fibrillation, and basic complications as significant factors influencing time to death at a 5% significance level. Furthermore, clustering effect between hospitals appears to have a significant impact on the time it takes for stroke patients to die. This suggests that the presence of frailty (clustering) effects underscores the importance of considering hospital-level heterogeneity in understanding the time-to-death of stroke patients. In light of these findings, the recommendation of the study emphasizes the importance of better hospital management, investigating additional factors, and effectively managing conditions such as hypertension, cardiac disease, diabetes mellitus, atrial fibrillation, and basic complications to improve outcomes for stroke patients. 77 Sat, 01 Jun 2024 00:00:00 GMT http://ir.haramaya.edu.et//hru/handle/123456789/8273 2024-06-01T00:00:00Z http://ir.haramaya.edu.et//hru/handle/123456789/8138 AUTOMATIC ANAPHORA RESOLUTION FOR AFAAN OROMOO LANGUAGE Elias Debelo; Dr. Wondwossen Mulugeta; Akubazgi Gebremariam Anaphora Resolution is a process of finding an entity introduced earlier in the discourse referred to by current entity back in discourse. The referenced entity is called antecedent, of referring entity which is called anaphor. There are number of anaphora types in a text, pronominal anaphora is the most prevalent one. Anaphora resolution is an important subtask that can be used in many Natural Language Processing applications. This study aims at developing Anaphora Resolution model based on Machine Learning approach for Afaan Oromoo language. The language is morphologically complex in that pronoun itself exist hidden inside the verbs. Input data cleaning, tokenization and part of speech tagging, noun phrase extraction and hidden pronoun extraction are useful steps toward Anaphora resolution. From every valid antecedent-anaphor pairs in training and testing sets, feature vectors will be generated. Machine Learning classifier trained using positive and negative instances generated from training set. Sklearn python package was used for as a trainer using fit function and as a predictor using predict function. Sklearn is set of packages consisting of implementations of Machine Learning algorithms. Three types of dataset, gathered from Afaan Oromoo News, Bible verses and Oromo Fictions, were used for training and testing. Five top best features were used for training and testing out of 14 features extracted from the text. Using 10-fold cross validation technique, the three datasets were divided into 10% testing and 90% training at each run. Each test sets of datasets were tested by range of 1 to 10 sentence distance between antecedent and anaphor on three Machine Learning algorithm Decision Tree (DT), Support Vector Machine (SVM) and Naïve Bayes (NB). Performance of the models on the three datasets were represented as mean average of the 10-folds on ten sentence range. Generally, average precision achieved for Bible-Fiction dataset 52.3% on DT, 51.25% using SVM and 53.57% using NB, for News dataset 57.67% using DT, 47.77% using SVM and 57.5% using NB and for compiled dataset 47.62% using DT, 46.82% using SVM and 50.15% using NB was achieved for combined independent and hidden anaphors. This result could be enhanced primarily by finding better ways of getting feature values for antecedent-anaphor pairs 96p. Thu, 01 Aug 2019 00:00:00 GMT http://ir.haramaya.edu.et//hru/handle/123456789/8138 2019-08-01T00:00:00Z