FDRE TVTI ETD
FDRE TVTI institutional repository is a new digital institutional repository system that collects, preserve and distribute the scholarly output of the Institute, mainly postgraduate Electronic Thesis and Dissertation (ETD), articles, proceedings. The system is dedicated to help users to find all the information they might require in order to format and successfully submit their graduate thesis, dissertation and publications electronically. The user friendly web interface enables to maximize and optimize the resource sharing among FDRE TVTI different faculties

Browse ETD Community
Select a community to browse its collections.
Recent Submissions
Housing Price Prediction Using Machine Learning v Case of Wolkite Town
(2026) Tseganew Kefie
This research focuses on developing a machine learning-based housing price prediction model for
Wolkite Town, Ethiopia. The existing studies on housing price prediction in Ethiopia are limited
and have not sufficiently considered local housing characteristics or ensemble learning techniques.
The dataset used in this study was collected from Wolkite Town housing records and contains
5170 housing instances with 22 housing features. The research methodology follows several
stages, including data collection and preprocessing, feature selection, training the model,
optimizing hyperparameters, and testing the model. After splitting the data into training and testing
sets, we applied Min-Max normalization to improve consistency. Several machine learning models
were implemented, including ElasticNet Regression, Random Forest Regressor, XGBoost
Regressor, LightGBM Regressor, and CatBoost Regressor. In addition, a Weighted Averaging
Ensemble model was developed by combining the two best-performing models, LightGBM and
Random Forest. Model performance was evaluated using Mean Absolute Error (MAE), Mean
Squared Error (MSE), Root Mean Squared Error (RMSE), and the R2
score. The results indicated
that the weighted ensemble model achieved the best performance, with an R2
score of 0.9396,
demonstrating high prediction accuracy. The findings show that ensemble machine learning
techniques effectively capture the housing market patterns and produce reliable house price
predictions. The developed model serves as a decision-support tool for home sellers, buyers, and
policymakers involved in housing-related planning in Wolkite Town.
Keywords: Machine Learning, Model, Housing Price Prediction, XGBoost Regressor,
Light
Predictive Model For Identifying High Milk-Yielding Sheko Cattle Cows Using Machine Learning Techniques In Sheka Zone, Ethiopia
(2026) Sultan Ahmed Hussen
This study developed and compared Decision Tree, Random Forest, XGBoost, and
AdaBoost machine learning models to classify milk yield in Sheko cattle. Data collected
from Teppi Agricultural Research Center included age, weight, body condition score,
parity, lactation stage, season, month, and health status. A quantitative research
methodology using supervised machine learning was applied. Models were developed in
Python using Google Colab with Scikit-learn, XGBoost, SHAP, Pandas, NumPy, and
Matplotlib. Data preprocessing included cleaning, label encoding, feature scaling,
SMOTE-based class balancing, and an 80:20 train-test split. Performance was evaluated
using accuracy, precision, recall, F1-score, and ROC-AUC. Random Forest achieved the
highest performance with 90.99% accuracy and a ROC-AUC of 0.9668. XGBoost was
used to validate model interpretation. SHAP analysis identified Health Status and Body
Condition Score as the most influential predictors, followed by Lactation Stage,
Design And Architectural Analysis Of A Scalable Cloud-Based Learning Management System For Remote And Blended Learning In Ethiopian Technical And Vocational Training
(2026) Sintayehu Feleke
LMSs have been essential platforms for remote and blended learning. Many LMSs are designed
for a highly stable and available internet connection, an adequate digital environment, and a
content-based learning paradigm. But, many LMSs do not consider the broader digital ecosystem,
infrastructural circumstances, heterogeneous digital device range, and competency learning model
of Ethiopian TVT institutions. This study aimed to design and analyze a scalable cloud-based LMS
architecture suitable for Ethiopian TVT institutions. The methodology of this study followed the
Design Science Research approach. It followed several steps, including conducting a literature
review, analysis of the existing situation, derivation of requirements, designing the architecture,
and conducting an analytical evaluation. The study proposed a context-aware cloud-based LMS
architecture reflecting infrastructural, operational, and educational challenges. The architecture
incorporates cloud-native design, offline-first synchronization, asynchronous processes,
bandwidth-adaptive content delivery, modular services, and competency-based assessment. Based
on the proposed architecture, a high-level system model and prototype model were designed to
reflect institutional processes, organizational structure, student progress and performance,
competency-based assessment, and blended learning support. The architecture was evaluated
through a traceability analysis, scenario-based analysis, architectural evaluation, and comparative
analysis.
Blockchain-Based Educational Certification System: A Case of TVET Colleges in Eastern Tigray
(2026) Shishay Hadush
An emerging trend towards faked certificates became a problem to educational integrity worldwide .
Eastern Tigray was known to be exposed to threats related to the conflict in that region, including
destruction of credentials and counterfeiting of certificates Traditional methods of verification were
deemed not sufficient as they posed a risk of both fraud and time-consuming process. It made the
process of credentials' verification difficult to manage for everyone. In this context, blockchain
technology offered a unique opportunity by providing an additional layer of protection due to the
decentralized and encrypted ledger. As a consequence, this paper explores the application of blockchain
technology to enhance the process of credential management in TVET colleges. Specifically, this
research will discuss ways of preventing fraud and destruction, rapid access to verification data, and
additional control over personal information that can be gained through the implementation of
blockchain-based credential management solutions. Methodology used in this research is mixed and
consists of qualitative interviews with TVET college’s administration, IT specialists, and students, as
well as quantitative analysis relying on the prototype of blockchain created using smart contracts
Classification Of Goat And Sheep Skin Diseases Using Deep Learning Approaches
(2026) Hiluf Nuguse Gidey
Animal farm has played one of critical roles in the socio economic wellbeing of most developing
countries, Ethiopia being one of them as sheep and goats play major role in providing food security,
generate incomes and export income. Nevertheless, there are a number of skin diseases, which
have a severe impact on the quality, and yields of animal products. In the rural set ups, they are
not handled well because of the inaccessibility of veterinary clinics and shortage of expertise in
the field. In this paper, an enhanced method of deep learning approach provided to identify and
classify skin diseases in sheep and goats with deep learning. The system automation CNN
architectures mean that features are automatically trained by skin images by which skin diseases
can be classified and classified to either bacteria and virus, parasite or healthy. The evaluated and
tested model underwent testing and evaluation on the standard measures, which include accuracy,
precision, recall and F1-score. The data employed in this thesis is 2341 images of data. The training
dataset classified as 80% (1873 images), 10% (234 images) and 10% (234 images) in training,
validation and test performance respectively. The classification work will consist of four categories
namely bacterial, viral, parasitic, and healthy skin conditions. This study compares the
performance of VGG19, Custom CNN, DenseNet201, and InceptionV3 for image classification
using accuracy, loss, and training time. Results show that DenseNet201 achieved the best overall
