DSpace Repository

Communities in NWU Repository

Select a community to browse its collections.

Now showing 1 - 5 of 20

Recent Submissions

Exploring Grade 12 teachers’ pedagogical content knowledge of Similarity in Euclidean geometry
(North-West University, 2024) Shoba, G.N; Nkhoma, P.; Kunene, N.
Studies show that Euclidean geometry is a section of school mathematics that takes more work to learn and teach in South Africa. Learners need to do better in Grade 12 matric examinations. The role of teachers in the learning and teaching process is critical for learners' success. This study explored Grade 12 teachers' Pedagogical Content Knowledge (PCK) about similarity in Euclidean geometry, as their depth in a subject specific PCK influences their classroom practices. Issues raised in the Grade 12 NSC diagnostic reports of the five consecutive years, 2017 to 2021, formed a basis for what to pursue. A qualitative research method was followed to establish the teachers' PCK via their ability to identify and correct the misconceptions and errors of the Grade 12 learners using written work of similarity concepts in Euclidean geometry. Eight secondary school teachers who offer mathematics in the Gauteng North district in the Gauteng province were purposively sampled and participated in the study. The study's findings revealed that with some of the teachers, their PCK depth of similarity in Euclidean geometry is lacking in identifying and addressing learner misconceptions and errors. Their challenges are similar to those of learners, as reported in the NSC diagnostic reports. These findings have several implications. This highlights the greater challenge of breaking the cycle of learners' poor performance in Euclidean geometry in South African schools. Secondly, the study's outcomes identify where one can intervene in improving the teachers' PCK. Lastly, it poses the challenge of revisiting the school curriculum by the Department of Basic Education and for South African tertiary institutions that train secondary school teachers to also re-visit their training programmes.
Die toepassing van 'n nifedipien-2-hidroksiepropiel-β-siklodekstrien insluitingskompleks vir die ontwikkeling van parenterale doseervorme
(North-West University, 1994) Pienaar, Heindrich; van Wyk, C.J.; Kotze, A.F.
Die siklodekstriene en hulle derivate het unieke ringvormige strukture wat hulle geskik maak vir die vorming van insluitingskomplekse met 'n verskeidenheid geneesmiddels. Hulle word veral in die farmasie gebruik om die biobeskikbaarheid, oplosbaarheid en stabiliteit van geneesmiddels te verhoog. Die derivate van beta-siklodekstrien, veral 2- hidroksiepropiel-beta-siklodekstrien, beskik oor beter oplosbaarheids-eienskappe as die natuurlike siklodekstriene en is hierbenewens ook minder nefrotoksies en hemolities. Hulle hou dus groot potensiaal in vir die formulering van geneesmiddels in parenterale doseervorme. Verskeie metodes vir die bereiding van geneesmiddelsiklodekstrienkomplekse word in die literatuur beskryf, waaronder die vloeistofvloeistofmetode die geskikste metode blyk te wees. Vries- en sproeidroging word algemeen gebruik om geneesmiddel-siklodekstrienkomplekse te isoleer vir identifikasie, analise en heroplossing. Verskeie metodes word ook gebruik vir die bestudering en beskrywing van die eienskappe van geneesmiddel-siklodekstrienkomplekse waaronder X-straalpoeierdiffraksie, fase-oplosbaarheidsdiagramme, kernmagnetiese resonansspektrometrie, en differensiele skanderingskalorimetrie die bekendste is. Nifedipien, 'n kalsiumkanaal antagonis, is 'n tipiese geneesmiddel wat swak wateroplosbaar is, en vinnige fotochemiese afbraak ondergaan. Hierdie eienskappe van nifedipien maak dit prakties onmoontlik om 'n stabiele waterige oplossing van nifedipien vir inspuiting te formuleer. Die voordele wat 'n parenterale doseervorm kan inhou, is van groot waarde veral ten opsigte van die biobeskikbaarheid en tempo van werking en die verspreiding daarvan in gehospitaliseerde pasiente. In hierdie ondersoek is daar 'n insluitingskompleks van nifedipien met 2- hidroksiepropiel-beta-siklodekstrien berei, en die invloed daarvan op die oplosbaarheid en stabiliteit van nifedipien ondersoek. Uit die ondersoek blyk dit dat die wateroplosbaarheid van nifedipien dramaties verhoog is, naamlik van ongeveer 6, 1 tot 261,4 µg/ml (0,0176 M tot 0,7548 M) by 'n 0,3846 M 2-hidroksiepropiel-betasiklodekstrienkonsentrasie. Op grond van hierdie resultate is dit wel moontlik om 'n waterige inspuiting met 'n dosis van 1 mg/5 ml te formuleer indien dit teen fotochemiese afbraak gestabiliseer kan word. 'n Stabiliteitsondersoek in deursigtige en amberglasampulle, onder verskillende ligtoestande is uitgevoer. Daar is gevind dat 2-hidroksiepropiel-beta-siklodekstrien nie die stabiliteit van nifedipien in deursigtige glasampulle vookom. Veranderinge in die ultravioletspektraaleienskappe, massaspektra, infrarooispektra, DSC-termogramme en X-straal poeierdiffraksiepatrone van nifedipien is gebruik om die insluitingskompleks van nifedipien met 2-hidroksiepropiel-beta-siklodekstrien te beskryf en aan te toon. Die resultate van hierdie ondersoek, tesame met die dramatiese verhoging in die wateroplosbaarheid van nifedipien na kompleksering met 2- hidroksiepropiel-beta-siklodekstrien toon duidelik aan dat molekulere interaksie tussen nifedipien en 2 hidroskiepropiel-beta-siklodekstrien bestaan.
Data Pruning: Redundant, Problematic, and Interdependent Samples
(2026) Freese, Leon; Marthinus, W; Theunissen
The performance of deep learning models is affected by not only data quantity but also data quality. Data pruning is a process by which practitioners can reduce the size of a dataset by only keeping the most important training data points, thereby achieving similar test set performance. We empirically investigate two popular data pruning methods under noisy and noiseless conditions and show that these methods fail in the presence of significant label noise. We highlight that the success of data pruning is distinctly affected by three factors: redundancy in the dataset, the presence of problematic samples, and interdependence between samples. We perform a detailed investigation on commonly used benchmark classification datasets and neural network architectures. We find that our observa- tions are consistent across data distributions and training protocols
Feature extraction for plant growth estimation
(2025) Ngorima, SA; Helberg, ASJ; Davel, MH
Precision agriculture requires the estimation of plant growth stages in real-time. When the plant growth stage is known, the wastage of resources in cultivation, such as nutrients and water, is reduced as only the required resources need to be supplied. Plants at different growth stages, however, have similar morphological features, which can make autonomous growth stage estimation difficult. This paper presents two feature extraction methods for growth stage estimation: one that uses a bank of Gabor filters and morphological operations, and the other that uses pre-trained convolutional neural networks (CNNs) and trans- fer learning. We test these methods on a publicly available plant growth stage dataset ("bccr-segset") for two species, canola and radish, grown and captured under indoor conditions. The two proposed feature extrac- tion methods are compared, using support vector machines and boosted trees as classifiers. We find that both methods are suitable for real-time applications, and that CNN features outperform the hand-crafted fea- tures, both with regard to speed and accuracy. The best system (VGG-19 features, classified with a radial basis function support vector machine) obtained an accuracy of 98.4% for both species, processing an image in 0.08 seconds.
Useful nonrobust features are ubiquitous in biomedical images.
(2026) Mouton Coenraad; Rabe Randle; Koser Niklas; Hansen Christopher
We study whether deep networks for medical imaging learn useful nonrobust features - predictive input patterns that are not human interpretable and highly susceptible to small ad- versarial perturbations - and how these features impact test performance. We show that models trained only on non- robust features achieve well-above-chance accuracy across five MedMNIST classification tasks, confirming their predic- tive value in-distribution. Conversely, adversarially trained models that primarily rely on robust features sacrifice in- distribution accuracy but yield markedly better performance under controlled distribution shifts (MedMNIST-C). Overall, nonrobust features boost standard accuracy yet degrade out- of-distribution performance, revealing a practical robustness- accuracy trade-off in medical imaging classification tasks that should be tailored to the requirements of the deployment setting.