Sanitation practices in Nigeria are shaped by environmental factors such as the location of communities (urban, peri-urban, or rural) and the availability of water, pressures that are intensified in conflict-affected settings such as Borno State. Using a mixed-methods design, primary quantitative and qualitative data were collected from 439 households in Maiduguri Metropolitan Council (MMC, urban) and Jere Local Government Area (peri-urban/semi-rural), Borno State, and analysed using chi-square tests and thematic coding; secondary national data (WASH-NORM, JMP) provided broader comparative context. More MMC respondents used improved sanitation facilities than respondents in Jere, a difference that was statistically significant (χ² = 10.370, p = 0.016). Sanitation facility type was also significantly associated with water-source reliability: households with piped water or boreholes mostly used flush toilets, while those depending on hand-dug wells or water vendors mainly used pit latrines (χ² = 30.240, p < 0.001). Qualitative accounts further showed that seasonal flooding, unstable sandy soils, and long distances to facilities damaged latrines, caused pit collapse, and discouraged consistent use. Environmental and infrastructural conditions strongly influence the use and sustainability of sanitation facilities in Maiduguri and Jere. Sustainable sanitation in conflict-affected Borno State requires flood-resistant designs, integrated water-and-sanitation planning, and greater investment in peri-urban sanitation infrastructure.
The orientation of brake pad slots significantly influences frictional heat generation, cooling behavior, and thermal stability in automotive disc brake systems under transient braking conditions. In this study, the transient thermal response of disc brakes equipped with vertical and inclined slotted non-asbestos organic (NAO) brake pads was investigated using experimental measurements and finite element analysis (FEA). Disc surface temperatures were measured at multiple radial locations under identical operating conditions. The results show that the vertical-slotted pad produced a lower peak disc temperature (≈283°C) than the inclined-slotted pad (≈315°C), corresponding to an approximate 10% reduction, while the experimental–numerical deviation remained within 5%. Radial temperature analysis revealed more uniform and speed-independent thermal profiles for the vertical-slot configuration, whereas inclined slots promoted localized heat accumulation and steeper thermal gradients. Cooling analysis indicated that the higher cooling rates observed for inclined slots were mainly due to their higher operating temperatures rather than improved heat dissipation efficiency. Overall, the results demonstrate that vertical-slot brake pads provide superior transient thermal stability, lower peak temperatures, and improved resistance to thermal localization under severe braking conditions.
The Nigerian banking sector faces a critical and escalating challenge with financial fraud that threatens the sustainability and trustworthiness of electronic banking services. Despite significant investments in security infrastructure and fraud prevention systems, Nigerian banks continue to experience substantial financial losses due to fraudulent activities. Financial fraud in banking systems is sparse, adaptive, and costly to monitor with fixed rule sets. To this effect, this paper presents an unsupervised autoencoder framework that learns the patterns of legitimate transactions and identifies suspicious activity using the reconstruction error of a symmetric autoencoder. The empirical evaluation uses the PaySim mobile money dataset, which contains 6,362,620 transactions and a fraud rate of 0.13 percent. After preprocessing through data cleaning, normalisation, feature encoding, and dimensionality reduction, eight numerical features remain for model training. The symmetric autoencoder converges after 11 epochs, and the anomaly threshold is set at the 95th percentile of the training reconstruction-error distribution, 0.001114. On the held-out test set, the model attains precision of 1.00%, recall of 38.96%, F1-score of 1.95%, specificity of 95.01%, accuracy of 94.94%, and AUC-ROC of 0.7338. The results indicate a useful screening mechanism for settings where labels are scarce, noisy, or unavailable. The framework is most significant as a first-stage detector that supports analyst review rather than as a final autonomous decision engine
Polycystic Ovary Syndrome (PCOS) affects an estimated one in ten women of reproductive age and presents with a heterogeneous range of metabolic, endocrine, and reproductive symptoms, complicating consistent clinical diagnosis. Methodology: Two related PCOS datasets (2,000 and 541 samples) were cleaned and combined for model training and external validation. Recursive Feature Elimination with Cross-Validation was used for feature selection, and class imbalance was addressed using the Synthetic Minority Over-sampling Technique (SMOTE), applied within each cross-validation fold to avoid data leakage. Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), and a soft-voting ensemble of the three were trained and evaluated using accuracy, sensitivity, specificity, precision, F1-score, and ROC-AUC, with statistical significance assessed via McNemar's test. Results: On the external test set, SVM, RF, and the ensemble achieved comparable accuracy (96.7-97.2%), all significantly outperforming LR (90.5%; p < 0.0001), with no significant difference among the three (p > 0.05). SHAP and LIME were applied for global and local interpretability, identifying follicle count, weight gain, skin darkening, and hormonal markers (AMH, cycle regularity) as leading predictors, though follicle count's dominance also raises a diagnostic-circularity concern given its role as a clinical diagnostic criterion. Conclusion: This framework demonstrates that ensemble and tree-based models substantially outperform linear baselines for PCOS risk prediction, with the ensemble offering comparable predictive performance alongside interpretable, dual-level explanations. These findings contribute toward more transparent and clinically trustworthy machine learning tools for PCOS diagnosis support.
The increasing deployment of hybrid photovoltaic (PV) power systems have created a growing need for intelligent operational monitoring to improve system reliability and maintenance. However, conventional monitoring techniques often fail to distinguish operational faults from weather-induced performance variations. This study presents a data-driven framework for intelligent operational monitoring by integrating residual analysis, statistical validation, and unsupervised K-means clustering. The framework was validated using three years of operational and meteorological data (June 2022–May 2025) from a 20 kWp hybrid PV system at the University of Nigeria, Nsukka. Expected daily energy output was estimated from solar irradiance and installed capacity, while residual analysis quantified deviations between expected and measured performance. The proposed framework analysed 1,488 operating conditions, identifying 640 anomalous (43.0%) and 848 normal (57.0%) conditions. The K-means model achieved a silhouette score of 0.62, indicating good cluster separation, with anomaly and normal cluster centroid residuals of −35.21 kWh and −96.47 kWh, respectively. The analysis revealed persistent negative residuals, zero-generation periods, and inverter interruptions, demonstrating that performance losses were primarily operational rather than environmental. The proposed framework distinguished weather-induced variability from technical faults and provides a practical decision-support tool for intelligent operational monitoring, early fault detection, predictive maintenance, and improved reliability of hybrid photovoltaic power systems in tropical environments.
