Audrey Karungi Ngambeki

Ms. Audrey Karungi Ngambeki

orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 June 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
Audrey Karungi NgambekiAudrey Karungi Ngambeki
Audrey Karungi Ngambeki
CHChongomweru Halimu
Chongomweru Halimu
FKFlorence N. Kivunike
Florence N. Kivunike
Reputational risk embedded in digital narratives significantly influences stakeholder perceptions of megaprojects, yet computational approaches for its systematic assessment remain limited. This study proposes an explainable sentiment-informed approach that operationalizes reputational risk as a governance-oriented text classification task using weakly supervised discourse analysis. A corpus of 25,480 records was compiled from online news and social media sources related to Ugandan megaprojects and processed using natural language processing techniques. To assess label reliability, three independent annotators evaluated a subset of 2,000 records, yielding an average Cohen's kappa coefficient of 0.7622, indicating substantial agreement. A soft-voting ensemble comprising Random Forest, Gradient Boosting, Logistic Regression, and Support Vector Machine classifiers was developed and evaluated. The model achieved an accuracy of 92.37%, a weighted F1-score of 0.9217, and strong agreement with reference labels (Cohen's kappa > 0.80). SHAP-based explainability analysis revealed that governance-sensitive discourse features were the primary drivers of reputational risk predictions, with corruption-related language emerging as the strongest indicator of high-risk narratives. The findings demonstrate the potential of interpretable machine-learning approaches for monitoring reputational risk in governance contexts and supporting evidence-based decision-making in megaproject management.

21 January 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
Audrey Karungi NgambekiAudrey Karungi Ngambeki
Audrey Karungi Ngambeki
HCHalimu Chongomweru
Halimu Chongomweru
FNFlorence Nameere-Kivunike
Florence Nameere-Kivunike
Reputational risk in textual narratives is a vital aspect of understanding stakeholder perceptions of megaprojects; however, formal computational methods for measuring it remain underdeveloped. This study introduces a computational model for senti-ment-based reputational risk, defined as a feature-based supervised classification task. The proposed model combines sentiment polarity, polarity intensity, topic distribu-tions, content length, and structural textual features into a structured mathematical model, enabling systematic evaluation and reproducible predictions. Data were col-lected from online news and social media, then processed through cleaning, tokenisa-tion, lemmatisation, and sentiment annotation. An ensemble model, merging Random Forests, Gradient Boosting, Logistic Regression, and Support Vector Machines via soft voting, was trained and assessed using accuracy, precision, recall, F1-score, and Co-hen’s Kappa. Results suggest that reputational risk can be reliably inferred from the interaction between sentiment, topics, and textual structure. Analyses of feature im-portance highlight polarity intensity, risk scoring, content length, and topic distribu-tion as key predictors. These findings demonstrate the potential of formal computa-tional models to quantify and predict risk within textual data. Future research could expand this model with transformer models and multilingual datasets to improve context-aware insights, explainability, and scalability, thereby laying a foundation for generalised computational approaches to reputational risk modelling.

7 October 2026 in Information
Verified
Journal Article
Open access
PDF file
Audrey Karungi NgambekiAudrey Karungi Ngambeki
Audrey Karungi Ngambeki
CHChongomweru Halimu
Chongomweru Halimu
FKFlorence N. Kivunike
Florence N. Kivunike
Reputational risk reflected in digital public discourse can significantly influence stakeholder perceptions, public trust, and decision-making surrounding government megaprojects. Existing computational approaches primarily focus on sentiment classification, offering limited capability for transparent and explainable modelling of reputational risk in public sector governance. This study proposes an explainable sentiment-informed model for identifying textual indicators of discourse-based reputational risk in digital discourse surrounding Ugandan government megaprojects. A corpus of 25,000 publicly available records from online news and social media sources was processed using a natural language processing pipeline, with weakly supervised labels subsequently validated through human validation of 2000 records by three independent annotators. A Soft Voting Ensemble integrating Random Forest, Gradient Boosting, Logistic Regression, and Support Vector Machine classifiers was evaluated using stratified, independent human, and temporal validation. Under the primary stratified evaluation, the model achieved 96.0% accuracy, 0.9592 weighted F1, 0.9327 Macro-F1, and 0.9079 MCC. Independent human validation yielded 72.5% accuracy and 0.5952 Macro-F1, while temporal validation yielded 89.4% accuracy and 0.6108 Macro-F1. SHAP analysis of the Soft Voting Ensemble identified governance and project-related textual features contributing to model predictions. The findings demonstrate strong primary performance while also showing that performance is lower under independent human and temporal validation, supporting the use of the approach as an interpretable decision support framework for identifying discourse-based reputational risk signals rather than as a validated predictor of realized reputational outcomes.