The Pentecost University is a private university located at Sowutuom in the Greater Accra Region of Ghana. It was founded by The Church of Pentecost (COP) and evolved from The Pentecost Bible College which initially trained only Lay Leaders and full-time Ministers for the COP. On May 22, 2003, J. A. Kufuor, the former President of Ghana, inaugurated PUC at the Sowutuom campus. The first PUC Council was inducted on May 6, 2004. Pentvars was accredited by the National Accreditation Board (NAB), Ghana in November 2004 and awarded its Presidential Charter on May 28, 2020, by Nana Addo Dankwa Akufo-Addo, the President of the Republic of Ghana. Prior to receiving a Presidential Charter, the university was affiliated to the Kwame Nkrumah University of Science and Technology, University of Cape Coast, and University of Ghana. On June 1, 2020, the university announced the appointment of Rev. Prof. Kwabena Agyapong-Kodua, who replaces Apostle Daniel Okyere Walker, as the first Vice-Chancellor since the university became fully-fledged.The student population is about 3000, made up of students enrolled in degree, certificate and professional programmes.The main campus of Pentecost University is sited at Sowutuom, away from Accra. Sowutuom (which literally translates “hold your gun”) is within the western parts of Accra, a quiet and serene area off the Kwashieman-Ofankor highway..
This study sought to ascertain whether Hofstede’s cultural dimensions are good predictors of cyberbullying behaviour among Ghanaian tertiary students, given that cultural values can shape individuals’ perceptions and actions, including what is considered acceptable during conflicts in digital spaces. This study used a quantitative, cross-sectional survey design to collect data from 301 tertiary students in Ghana. The data was analysed using structural equation modelling. The model explained 22.9
The study aims to enhance African students’ understanding of cybersecurity by integrating African culture into higher education. It uses the Culturo-Techno-Contextual Approach (CTCA), a pedagogical method based on using indigenous knowledge to break down complex scientific concepts, and the Afrocyberlibrary, a database of Indigenous knowledge in cybersecurity. The quantitative phase involved a quasi-experiment with 71 students, while the qualitative phase had seven participants. The study showed that CTCA and Afrocyberlibrary improved student achievement in cybersecurity compared to the regular lecture method; its effects on gender were inconclusive. The CTCA was praised for blending culture and technology, making cybersecurity learning more meaningful, and employing cultural examples to explain concepts. Students appreciated the intuitive interface and rich educational content of the Afrocyberlibrary. This study’s findings imply that Afrocyberlibrary’s combination of indigenous knowledge and its application to contemporary digital education can contribute to the digitisation and democratisation of higher education.
Thermal drift remains a major source of geometric inaccuracy in CNC machine tools, especially during long-horizon operation as temporal regimes evolve. This paper presents ThermaForge, a learned adaptive thermal error compensation policy for processed thermal-forecasting settings in which explicit machine-state branches are unavailable. Instead of relying on a single end-to-end sequence model, ThermaForge combines chronological split-aware base prediction (ridge and linear models) with persistence-aware residual meta-features and a calibrated policy head. Experiments on the Mendeley three-axis thermal error benchmark under strict chronological evaluation show consistent gains over classical and sequence baselines. On the main $w=30$ profile, ThermaForge achieves the best root mean square error (RMSE) (0.469250) and coefficient of determination $\left(R^{2}\right)$ (0.115164), outperforming the strongest baseline (ridge; RMSE 0.469682) and ranking first for all five seeds. In the $w=60$ ablation regime, the calibrated ThermaForge variants reduce RMSE by about 0.00356 relative to persistence while improving $R^{2}$ from 0.099647 to about 0.113. On the live ZK5540 follow-up dataset, ThermaForge attains an RMSE of 0.011375, compared with 0.029168 for ridge and 0.029961 for linear regression, demonstrating strong performance on real-machine data. These results indicate that learned policy-level calibration, coupled with leakage-safe temporal training, improves the robustness of thermal compensation-oriented prediction.
Integrating technology into teaching is essential to achieving equitable, high-quality education, particularly within the framework of Sustainable Development Goal 4. While prior research in Ghana has largely examined Information and Communication Technology (ICT) use among in-service teachers and at the senior high school level, limited attention has been given to pre-service teachers and the role of gender in shaping technological dispositions. The growing emphasis on technology integration in mathematics requires teacher preparation programs to adequately prepare future teachers with dispositions that support meaningful technology integration. Guided by the Theory of Planned Behavior, this study examined pre-service mathematics teachers’ attitude or disposition towards technology integration in mathematics instruction and investigated whether this attitude differs by gender. The article reports the quantitative component of a larger sequential explanatory mixed-method study. This study examined the perspectives of 301 pre-service mathematics educators from five Colleges of Education in the Ashanti region of Ghana regarding their attitudes towards ICT integration in mathematics teaching. Data were collected from 301 participants using a structured questionnaire adapted from existing technology-related instruments and were analyzed using mean, standard deviation, and an independent t-test. The findings indicated predominantly unfavorable attitudes or dispositions towards ICT integration in mathematics, with participants perceiving technology as frustrating, time-consuming, and unnecessarily complicating mathematics. A statistically significant gender difference was observed, with male pre-service teachers having notably more positive attitudes than their female counterparts, t (299) = 2.94, p = .004, Cohen’s d = 0.34. These findings suggest that unfavorable technological dispositions may constrain pre-service teachers’ readiness to integrate technology meaningfully into mathematics instruction. The findings of this study indicate the necessity for teacher education programs to incorporate mathematics-specific technology pedagogy, tackle gender disparities in ICT confidence and engagement, and enhance institutional support for technology integration in the classroom. The study highlights critical gaps in teacher preparation for ICT integration in resource-constrained contexts.
In recent years, the research of leveraging deep learning algorithm to develop thermal error compensation which is the main factor that affect machining accuracy and thermal error caused by thermal deformation. The improvement of modern machine tool manufacturing technology has most significantly influenced the accuracy and energy efficiency of machining tool quality. Traditional machine learning modeling methods of thermal error relied on the selection of temperature sensors location and have a lack of prediction data validation. However, in actual production the working conditions are complicated and it is relatively easy to collect the data of the rest and idling states, while the data collection under the processing state is difficult. To solve these problems, this paper proposes a new modeling method based on generative adversarial network which predict thermal error before the beginning of machining. This method generates the thermal error in every moment of machining, both for idling with constant speed and milling. The method does not rely on troubling selection of many temperature sensors or complicated mathematical inference. This method was easy to be implemented, which was able to achieve energy savings and smooth compensation. Experimental results showed that the method reached an accuracy of 91% for idling prediction and 70% for milling prediction. In conclusion the case of insufficient data volume, the proposed method was superior to traditional machine learning methods in terms of prediction accuracy and model training efficiency.
