Social media adoption over the years, has seen billions of people participating on social networking sites and the interaction thereof has considerably influenced marketing related correspondence, public relations as well as modern day business communication. This new level of marketing communication that is evolving yearly brought forth new opportunities and challenges for companies as purchase decisions are increasingly being influenced by social media interactions as consumers rely more than ever on their social media interaction when making those decisions. The main purpose of this study was to examine the influence of social media on brand awareness, the effect of brand awareness on perceived quality, brand loyalty and purchase intention, effect of brand loyalty on purchase intention and finally the effect of perceived quality on brand loyalty and purchase intention. To attain the empirical objectives of the study, a quantitative research design was adopted and the respondents were selected using a non-probability convenience sampling technique with data collection done using self-administered structured questionnaire to 320 respondents. The Statistical Package for Social Sciences (SPSS) version 27.0 was used to compute both descriptive and inferential statistics (which includes correlation and regression analysis). In addition, the reliability and validity analysis were carried out in relation to the measuring instrument. Through regression analysis, positive significant predictive relationships were found between social media interaction and brand awareness, brand awareness and perceived quality, brand loyalty and purchase intention, brand loyalty and purchase intention, perceived quality with brand loyalty and purchase intention. Furthermore, the results of the study also pointed out that social media interaction has influence over the consumer purchase decision process. In light of the growing influence of social media, organisations should invest in creating customised interactive social media campaigns targeting loyal consumers with the objective of enhancing the behavioural motivation to 'Like' in order to engage WOM, as it can more effectively generate a higher purchase intention.

This study investigated postgraduate students’ perceptions of research quality within a higher education context, focusing on a selected University of Technology (UoT) in Gauteng South Africa. Drawing on a mixed-methods research design, the study integrated quantitative and qualitative data to examine how institutional and individual factors shape research quality and outcomes. Data was collected through a structured questionnaire and interview responses from registered master’s and doctoral students. Quantitative findings, analysed using descriptive statistics, indicate that supervision quality and research training are positively perceived and play a significant role in enhancing students’ research capacity. However, institutional support and the broader research environment were rated moderately, suggesting the need for improved access to research resources and more effective institutional support systems. Funding also emerged as a critical determinant of research quality, with participants highlighting the impact of financial constraints on research progress. The study contributes to the growing discourse on postgraduate studies by highlighting the need for a holistic approach to research quality. It recommends strengthening supervision practices, increasing funding support, and enhancing institutional resources to improve quality research. Existing literature indicates that these funding challenges are deeply rooted in systemic and historical inequalities that continue to shape higher education institutions in South Africa, despite successive policy reforms aimed at improving quality research and student access. These persistent disparities not only affect students’ ability to access quality postgraduate studies but also influence their academic progression, research quality, and overall success. These research study findings have significant contribution for policy development and practice in higher education institutions seeking to promote quality research.

High-throughput docking generates candidate poses faster than affinity can be measured experimentally. We encode binding pockets and ligand graphs with message-passing layers trained on curated protein–ligand complexes and hold out scaffolds unseen during optimisation. The model ranks actives above decoys with area under the curve comparable to physics-based rescoring on two kinase panels while requiring a fraction of the compute, suggesting utility for prioritising medicinal chemistry series before synthesis.

Seed diffusion depends on neighbour imitation, credit access, and risk preferences that aggregate surveys often collapse. We parameterise an agent-based model with plot-level trial data and focus-group elicitation of social learning rules. Simulations show that modest subsidy on first-season seed bundles accelerates coverage more than price discounts alone when extension agents coordinate demonstration plots along existing kinship networks rather than along administrative boundaries.

Mechanical ventilation upgrades are unevenly funded across public school districts. We deploy low-cost carbon dioxide loggers in twenty-four classrooms over two academic terms and link weekly exceedance minutes to anonymised absence records. Elevated afternoon CO2 concentrations correlate with modest increases in respiratory-related absences after adjusting for outdoor particulate levels and heating season, supporting targeted window-management protocols and filter maintenance schedules where capital retrofits remain deferred.

Boat-based rake surveys are labour intensive and miss patchy beds near turbid inflows. We acquire multispectral orthomosaics at low tide windows and train random forest classifiers on in situ depth and species notes. Mapped bed extent agrees with diver transects within ten percent on clear days, enabling seasonal tracking of regrowth after mechanical harvesting and informing buffer strips that reduce nutrient loads from adjacent cropland.

Adaptive comfort models assume occupants can modify clothing and airflow, yet many public-sector offices restrict window opening. We combine wearable skin-temperature proxies, subjective votes, and keystroke-level task metrics during summer weeks. Predicted mean vote bands derived from adaptive equations overestimate acceptability when afternoon humidity spikes coincide with fixed dress codes, implying that retrofit prioritisation should weight dehumidification alongside shading for knowledge-work floors.

Coastal and highland destinations respond differently to currency movements and charter-flight capacity. We estimate a panel vector autoregression on monthly receipts, nominal exchange rates, and hospitality payroll for six regions with heterogeneous seasonality. Impulse responses indicate that depreciation shocks raise dollar-denominated receipts with a lag in lake circuits linked to cross-border overland travel, whereas beach corridors react faster through prepaid international packages.

Structured reporting forms capture only a fraction of reactions documented in discharge summaries. We fine-tune a transformer-based semantic role labeler on annotated sentences linking drugs, dosages, and reaction phrases, then apply it to a de-identified corpus from regional hospitals. Precision for serious events improves over keyword baselines, and error analysis highlights negation scope and combination therapies as remaining challenges for safety signal triage.

Guidelines recommend multimodal care, but trials compare heterogeneous packages with sparse head-to-head evidence. We synthesise randomised studies of structured exercise, manual therapy, and brief cognitive behavioural programmes using a network meta-analysis with arm-level pain and disability scores at twelve weeks. Combined exercise and education ranks highest for function gains, while manual therapy alone offers short-term relief without sustained advantage, supporting stepped-care algorithms feasible in nurse-led clinics.