Ounongo Repository
The Ounongo Repository (OR) is the institutional repository of Namibia University of Science and Technology. Ounongo means "knowledge. in the Oshiwambo and Otjiherero languages. The OR is administered by the Library, with technical assistance from DICT, and its aim is to collect, organize, manage, store, preserve, publish and make accessible worldwide, the knowledge assets or intellectual output of the University's researchers, staff and post-graduate students. Users may set up RSS feeds to be alerted to new content.
Communities in Ounongo
Select a community to browse its collections.
Recent Submissions
Co-Designing And Implementing Independent Journalism And Archiving With The Indigenous San Community In Donkerbos Through A Self-Sustainable Model.
(Namibia University of Science and Technology, 2024-08-15) Kaulbach, Peter
This master thesis explores the impact of a multimedia project in a San community, focusing on podcast episodes and digital storytelling initiatives. The project aimed to amplify the voices and stories of the San people [in Donkerbos] while addressing various topics such as politics, cultural practices, and community projects. Through participatory methodologies, the project engaged community members in the production process, ensuring cultural sensitivity and authenticity. The study responds to the systemic underrepresentation and misrepresentation of San communities in mainstream media, which are often shaped by external perspectives. It investigates the project’s influence on the San community’s awareness, empowerment, cultural preservation, and socio-economic development. It examines the reception of podcast episodes among community members, accounting for their perception and attitude toward the content. Additionally, the thesis explores the effectiveness of digital storytelling initiatives, such as videos on social media platforms, in promoting cultural awareness and challenging stereotypes. Methodologically, the thesis employs a mixed-methods approach, including focus groups, workshops, and content analysis. It draws on theories of indigenous media, participatory communication, and cultural preservation to frame its analysis. The findings show that participants gained technical media skills, expressed increased cultural confidence, and began engaging more actively in the documentation of their community’s stories. Challenges included infrastructural limitations and questions of long-term sustainability. The research contributes to the growing body of literature on indigenous media, community-based research, and cultural preservation. It concludes that participatory multimedia initiatives can enhance cultural resilience and representation, while offering a sustainable framework for community-driven storytelling.
Forecasting The Consumer Price Index in Namibia: A Comparative Analysis of Machine Learning and Statistical Methods
(Namibia University of Science and Technology, 2025-10-02) Elago, Linea.
In emerging countries like Namibia, accurate forecasting of the Consumer Price Index (CPI) is important for evidence-based policy development, effective economic planning, and inflation management. However, despite its importance, there has been limited application of advanced forecasting methods within Namibia’s context. The study addressed this gap by investigating and comparing the performance of traditional statistical methods, AutoRegressive Integrated Moving Average (ARIMA), Holt-Winters exponential smoothing with machine learning methods, Long Short-Term Memory (LSTM) recurrent neural network, and Support Vector Regression (SVR) in forecasting Namibia’s CPI using monthly data from 2013 to 2023. The findings revealed that SVR yielded the lowest Root Mean Square Error (RMSE), indicating higher forecast accuracy compared to other models. The study recommends a shift towards machine learning methods, particularly SVR, for CPI forecasting, given its capability to capture nonlinear trends in economic data. The study further recommends enhancing forecasting methods by incorporating relevant economic indicators such as interest rate, GDP growth, unemployment rate, and government expenditure. The study also emphasises the importance of investing in capacity-building for data science and developing real-time data systems to support policy formulation and CPI monitoring. The research contributes to the growing body of evidence that machine learning can enhance CPI forecasting and decision-making processes in developing countries
Language as an expression of anger in selected Namibian novels: Masked warrior and Complicated
(Journal of Communication and Cultural Trends, 2025-03-07) Kambwale, Elizabeth Ndavavaelao; Woldemariam, Haileleul Zeleke
This article presents a cognitive stylistic study of anger in two Namibian novels: Ndinaelao Moses’ Masked warrior and Malakia Haimbangu’s Complicated. The study evaluated the lexical expressions of anger, figurative expressions, and features of anger discourse. The study aims at probing on how the language in the selected Namibian fictional works deals with anger expressions, particulary on how authors represent societal problems through a cognitive stylistics approach. The study applied textual world theory as a theoretical framework for understanding and analysing the texts. It follows a qualitative approach, with content analysis as the primary data collection method. The results of the study showed that the texts had manipulated and maintained the readers' interest through the use of anger. The study found that words about anger are made more offensive by using figurative language terms. Additionally, the study showed that angry language might be used to show defensiveness, sorrow, or arrogance. The study found that creating writings with anger in them makes readers relate to the characters’ real-world experiences. The findings further established a key communicative function of figurative language that is simplification. Specifically, the study concluded that the strategic deployment of figurative phrases facilitates the effective transmission of complex or abstract concepts, thereby augmenting their accessibility and comprehension for the target audience. The study concluded that discourse influences how angry texts are written. The study suggests the use of alternative language and grammatical expressions that are consistent with Text World Theory, which emphasises the significance of using linguistic and cognitive strategies to create a cohesive and immersive fictional world.
Predictive Modelling of Taxpayer Compliance Behaviour Using Machine Learning At NamRA
(Namibia University of Science and Technology, 2025-07-31) Sifani, Edwin Sifani
This study explored the application of machine learning techniques for predictive modelling of taxpayer compliance behaviour at the Namibia Revenue Agency (NamRA). Multiple classification algorithms were systematically optimised using 5-fold cross-validated Grid Search and Randomised Search, as implemented in the scikit-learn library (v1.2), to enhance predictive accuracy. The hyperparameter search spaces were tailored to each model’s architecture; for instance, Random Forest optimisation included the number of estimators and maximum depth, while Gradient Boosting models emphasised learning rate and structural parameters. The optimisation process yielded notable improvements, with cross-validated accuracy scores ranging from 64% to 68%. The best-performing model, an optimised Random Forest classifier, achieved an accuracy of 68%. These findings demonstrated the efficacy of hyperparameter tuning in improving model performance and underscore the potential of machine learning to support data-driven compliance management at NamRA.
The use of SHAP and LIME in this study provided valuable interpretability of taxpayer compliance predictions, highlighting key factors such as income group, the COVID-19 period, taxpayer registration office, and marital status. These insights align with existing research and reveal how financial capacity, macroeconomic disruptions, and administrative or demographic variables influence compliance. SHAP offers a global view of feature importance, while LIME provides personalised explanations, enhancing trust and communication. Despite modest predictive accuracy, the interpretability benefits support targeted policy interventions and suggest future improvements through richer data and fairness assessments.
Developing A Data-Driven Financial Model for Decision Support in Evaluating Investment Portfolio Performance
(Namibia University of Science and Technology, 2025-09-12) Nghilundwa, Pendapala
Contemporary financial markets demonstrate heightened complexity and volatility, necessitating sophisticated instruments for the precise assessment of investment portfolios. This research explores the application of machine learning (ML) models to predict Month-to-Date (MTD) returns, aiming to enhance financial decision-making. Conventional models frequently exhibit insufficient flexibility to fluctuating market conditions, highlighting the necessity for data-driven approaches that prioritise portfolio-specific metrics, such as Market Value Dirty and Year-to-Date (YTD) returns. Meanwhile, macroeconomic variables such as gross domestic product (GDP) growth, inflation, and interest rates played a secondary role. The study employed a quantitative method, using secondary data from January 2017 to August 2024, which comprised financial measures and macroeconomic variables. Four machine learning models were developed, namely Random Forest, Gradient Boosting, XGBoost and Long Short-Term Memory (LSTM). Data preprocessing and feature engineering played a critical role in model development. Feature engineering involved creating cumulative MTD returns, moving averages (5-day and 10-day), and volatility metrics to capture market trends and risk dynamics. These features were derived from daily returns, grouped by year and month, and calculated using rolling operations. Data normalisation was applied to standardise input variables, and missing values resulting from rolling operations were filled to ensure dataset completeness. The dataset was then divided into training and testing datasets using a 1:1 ratio. Model performance was evaluated using mean squared error (MSE) and R-squared (R²) metrics, with cross-validation assuring robustness. Among the models, Gradient Boosting attained the lowest mean squared error (MSE: 2.39 × 10⁻⁶) and the highest R² (0.922), outperforming Random Forest (MSE: 2.72×10⁻⁶, R²: 0.911), XGBoost (MSE: 3.13×10⁻⁶, R²: 0.898), and LSTM (MSE: 3.98×10⁻⁶, R²: 0.870). Feature importance analysis highlighted Market Value Dirty, YTD Return, and Benchmark (BM) Size as the most influential predictors. At the same time, macroeconomic variables such as interest rates and inflation contributed minimally to short-term forecasting. This demonstrates the dominance of portfolio-specific metrics in predicting MTD returns. While LSTM excelled in capturing temporal relationships, its predictive accuracy lagged due to volatility during high-risk periods. The results affirm the effectiveness of machine learning models in enhancing financial decision-making. Gradient Boosting and Random Forest models offer accurate predictions and valuable insights into key portfolio-specific factors, underscoring their utility for risk management and strategic planning. The dissertation recommends further exploration of hybrid models, the inclusion of additional macroeconomic variables, and the integration of real-time data to enhance predictive accuracy and robustness. These improvements establish data-driven approaches as essential instruments for financial firms operating in unpredictable markets.