Claim Missing Document
Check
Articles

Found 5 Documents
Search

How Effective are Different Machine Learning Algorithms in Predicting Legal Outcomes in South Africa? Khosa, Joe; Mashao, Daniel; Olanipekun, Ayorinde; Harley, Charis
Journal of Applied Data Sciences Vol 5, No 4: DECEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i4.215

Abstract

This study examines the effectiveness of different machine learning algorithms in predicting legal outcomes in South Africa's Judiciary system. Considering the advancement of artificial intelligence in the legal sector, this research aims to assess the effectiveness of various machine learning algorithms within the legal domain. Text classification is done using machine learning algorithms, including Logistic Regression, Random Forest, and K-Nearest Neighbours, with datasets obtained from a state legal firm in South Africa. The datasets undergo diligent data cleansing and pre-processing methods, encompassing tokenization and lemmatization techniques. This study evaluates these models' applications through accuracy metrics. The findings demonstrate that the Logistic Regression model attained an accuracy rate of 75.05%, whereas the Random Forest algorithm achieved an accuracy rate of 75.08%. On the other hand, the K-Nearest Neighbours algorithm exhibited no optimal performance, as evidenced by its accuracy rate of 62.76%. This study provides valuable insights for legal professionals by addressing a specific research question about the successful application of machine learning in South Africa's legal sector. The results indicate the possibility of using machine learning to predict the outcomes of criminal legal cases. Additionally, this study highlights the significance of responsibly and ethically implementing machine learning within the legal field. The results of this study enhance our comprehension of the prediction of legal outcomes, establishing a foundation for future investigations in this dynamic area of study. A limitation of this study is that the data was obtained from a single law firm in South Africa.
Sentimental Analysis of Legal Aid Services: A Machine Learning Approach Khosa, Joe; Mashao, Daniel; Olanipekun, Ayorinde
Journal of Applied Data Sciences Vol 6, No 2: MAY 2025
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v6i2.521

Abstract

Legal Aid services in South Africa, administered by Legal Aid South Africa (SA), aim to provide essential legal representation to vulnerable individuals lacking financial resources. Despite its significant role, there is a pervasive perception among the public that the quality of these state-funded services is substandard, often leading to negative attitudes towards the organization. This research employs sentiment analysis to evaluate client perceptions of Legal Aid SA's services, using a dataset of 5,246 entries from Twitter and the Internal client feedback system between 2019 and 2024. The study utilizes various machine learning algorithms, including Naive Bayes, Stochastic Gradient Descent (SGD), Random Forest, Support Vector Classification (SVC), Logistic Regression, and Extreme Gradient Boosting (XGBoost), to analyze sentiment polarity and classify feedback into positive, neutral, and negative sentiments. The accuracy, precision, recall, and F1 scores assessed model performance. The SVC and XGBoost models demonstrated superior performance, achieving testing accuracies of 90.10% and 90.00%, respectively. In contrast, Naive Bayes and Logistic Regression lagged, with test accuracies of 82.00% and 85.00%, respectively. The findings reveal that most responses are either neutral or positive, suggesting a predominantly favourable impression of Legal Aid services. This research not only aims to enhance Legal Aid SA's service offerings but may also provide valuable insights for similar organizations globally.
Analyzing the Impact of Company Location, Size, and Remote Work on Entry-Level Salaries a Linear Regression Study Using Global Salary Data Khosa, Joe; Mashao, Daniel; Subekti, Fajar
International Journal of Informatics and Information Systems Vol 7, No 3: September 2024
Publisher : International Journal of Informatics and Information Systems

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/ijiis.v7i3.215

Abstract

This research explores the key factors influencing entry-level salaries in the global labor market of 2024, emphasizing the roles of company location, organizational size, and the extent of remote work in shaping compensation levels. Drawing on the Global Salary 2024 dataset from Kaggle, which comprises over 5,600 observations across multiple industries and geographic regions, the study applies a multiple linear regression model executed in Python via Google Colab to quantitatively examine salary disparities. The results indicate that company location and size significantly affect entry-level earnings, underscoring how regional economic contexts, cost-of-living variations, and organizational capacity continue to drive wage formation. Conversely, the remote work ratio exhibits a negligible and statistically insignificant effect, implying that flexibility in work arrangements has yet to translate into measurable financial value for early-career professionals. Furthermore, introducing job title as a control variable enhances the model’s explanatory power, reaffirming the influence of individual skill specialization and job function in determining compensation outcomes. These findings reinforce human capital theory while extending it by incorporating contextual and organizational dimensions relevant to the digital labor economy. For job seekers, the study offers data-driven insights to guide career decisions and salary expectations across regions, while employers may utilize the results to formulate fair and competitive pay strategies in an increasingly interconnected workforce. Ultimately, this study provides a comprehensive understanding of how structural and individual factors interact to shape entry-level salary dynamics in the modern digital era.
Analyzing GPU Efficiency in Cryptocurrency Mining: A Comparative Study Using K-Means Clustering on Algorithm Performance Metrics Khosa, Joe; Olanipekun, Ayorinde
Journal of Current Research in Blockchain Vol. 2 No. 2 (2025): Regular Issue June 2025
Publisher : Bright Institute

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jcrb.v2i2.34

Abstract

This study employs clustering analysis to evaluate the efficiency of GPUs used in cryptocurrency mining, categorizing them into distinct groups based on computational output and power consumption. Using K-Means clustering, GPUs were grouped into three clusters: low-efficiency, moderate-efficiency, and high-efficiency. High-efficiency GPUs demonstrated superior hash rates (e.g., 104.79 Mh/s for AbelHash and 218.35 Mh/s for Autolykos2) despite higher power consumption, making them ideal for high-performance mining operations. Conversely, low-efficiency GPUs exhibited lower computational output and modest energy use, highlighting opportunities for hardware upgrades or repurposing. Visualization techniques, including scatter plots and pair plots, provided clear distinctions between clusters, while a silhouette score of 0.35 indicated moderate cluster separation, suggesting areas for further refinement. The findings offer actionable insights for optimizing hardware selection, reducing operational costs, and improving energy efficiency in mining operations. Additionally, this study underscores the importance of sustainability in cryptocurrency mining and provides a foundation for future research, including the integration of additional performance metrics, exploration of alternative clustering algorithms, and development of energy-efficient mining practices. These insights contribute to the broader goal of fostering a more sustainable and data-driven approach to cryptocurrency mining.
Exploring the Impact of Mixed Reality Technology on Anatomy Education for Medical Students Khosa, Joe; Olanipekun, Ayorinde
International Journal Research on Metaverse Vol. 2 No. 1 (2025): Regular Issue March
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/ijrm.v2i1.18

Abstract

This study investigates the effectiveness of Apple Vision Pro, a mixed reality tool, in enhancing medical students' understanding of 3D anatomical structures compared to traditional teaching methods. A quasi-experimental design was employed, involving 500 medical students who were divided into two groups: the Experiment group (n = 250), which used Apple Vision Pro, and the Control group (n = 250), which relied on conventional 2D images, textbooks, and static models. Both groups completed a pre-test to assess baseline knowledge, followed by an intervention phase over three weeks, and a post-test to measure learning outcomes. The results showed that the Experiment group demonstrated significantly greater improvement in post-test scores, with a mean improvement of 19.56 ± 5.58, nearly double the 9.40 ± 2.86 improvement observed in the Control group. Statistical analysis using an independent t-test confirmed that this difference was highly significant (t = 36.20, p < 0.0001), indicating the superior effectiveness of Apple Vision Pro in facilitating spatial visualization and comprehension of anatomical relationships. Qualitative feedback from the Experiment group further highlighted the benefits of Apple Vision Pro, including its ability to deliver an immersive, interactive, and engaging learning experience. Students reported increased motivation and a deeper understanding of anatomical structures due to the dynamic nature of the mixed reality environment. In conclusion, this study provides compelling evidence that Apple Vision Pro can transform anatomy education by addressing limitations associated with traditional teaching methods. The findings suggest that integrating mixed reality tools into medical curricula can significantly enhance learning outcomes, improve student engagement, and foster a more comprehensive understanding of complex anatomical concepts. Future research should focus on evaluating the long-term impacts of mixed reality technologies on knowledge retention and practical skill development in medical education.