Browsing by Author "Muhaise, Hussein"

Now showing 1 - 4 of 4
  • Thumbnail Image
    ItemOpen Access
    Application Of ArtifiCial Intelligence in E-Governance: A Comparative Study of Supervised Machine Learning and Ensemble Learning Algorithms on Crime Prediction.
    (Kabale University, 2024) Niyonzima, Ivan; Muhaise, Hussein; Akankwasa, Aureri
    In the developing world, the daily activities of humans’ social, political and economic life make it vital and easy to encounter the phenomenon of crime. Crime is an unnecessary evil in society and for any economic, social and political activities to run smoothly, criminal offenses must be completely eliminated from society. Advancement in information and communications technology enables law enforcement agencies to collect a huge amount of crime data, and the data collected by these organizaions have been doubling every two years. It has been found out that only 17% of the collected crime data is used in their operations today and several studies have noted that Law Enforcement Agencies are data rich but information poor. Machine learning, a subfield of artificial intelligence, has been used by government agencies in developed countries in different operations like face recognition, computer forensics, image and video analysis to identify criminals and crime predictions. It is therefore time for developing countries to leverage such technologies in order to reduce crimes. Therefore, this study proposes the application of supervised machine learning techniques in the prediction of crimes basing on the past crime data. During this study, we used open-source crime data from the UCI Machine learning repository to train and validate our algorithms. The performance of supervised machine learning and ensemble learning algorithms was done using crime data. The supervised machine learning algorithms used include K-Nearest Neighbor (KNN), decision tree classifier (CART), Naïve Bayes (NB) and Support vector machine (SVM). The ensemble learning algorithms used include AdaBoost (AD), Gradient Boosting Classifier (GBM), Random Forest (RF) and Extra Trees (ET). We used an accuracy metric to measure the performance of the algorithms. Python 3 was used in all the experiments using windows 10 laptop with 8GB RAM and 2.0GHZ processor. The performance of the supervised machine learning algorithms using the original datasets includes 60.33%, 56.24%, 57.01% and 59.06% for KNN, CART, NB, and SVM respectively. The performance of ensemble learning algorithms using the original datasets includes 58.58%, 59.81%, 55.23% and 55.74% for AD, GBM, RF and ET respectively. Experimental results revealed that KNN generally performed better when compared to the rest of the algorithms. we then developed a crime prediction model based on KNN and its prediction accuracy was 66% on our test dataset. The use of Artificial Intelligence has the potential to ameliorate several existing structural inefficiencies in the discharge of governmental functions. Machine learning, a subfield of artificial intelligence, has been used by government agencies in developed countries in crime analysis and predictions. It is therefore time for developing countries to leverage such technologies in order to reduce crimes.
  • Thumbnail Image
    ItemOpen Access
    Assessing Medical Students’ Learning Style Preferences at Kabale University Medical School, Uganda.
    (Kabale University, 2024) Muhaise, Hussein; Businge, Phelix Mbabazi; Ssemaluulu, Paul; Kyomugisha, Patricia
    This article is based on an empirical study conducted to assess and establish the preferred learning styles of medical students in the Kabale University Medical School. The study was prompted by a paradigm shift in teaching-learning strategies from the conventional knowledge-based medical curriculum to competency-based medical education (CMBE). In line with the learners’ diversity and inclusion, CBME liberalizes the learning environment by providing a variety of learning methods. Hence, this study aimed to ascertain the preferences of medical students’ learning styles concerning competency-based learning approaches. Procedurally, the study employed online survey methods, and the respondents included 160 medical (MBChB) students, all from Kabale University School of Medicine. The data collected were captured on SPSS version 26 and subjected to t-test analysis. Besides, Visual, Aural, ReadWrite, and Kinaesthetic (VARK) learning inventory was used to determine the student’s learning preferences, while a t-test was used to establish the relationships between the demographic profiles and the learning styles. Notably, the Aural learning style produced the highest mean score of 7.21 ± 3.61, followed by Kinaesthetic (6.43 ± 3.22), ReadWrite (6.12 ± 2.23) and Visual (4.04 ± 2.42). Relatively, t-test results showed significant (p < 0.05) differences in learning styles between preclinical and clinical students. However, the t-test results for gender factors for all the learning dimensions were insignificant (p > 0.05). Pre-clinical students prefer visual and read-write learning styles, while clinical students prefer kinaesthetic and visual learning styles. Based on the findings, this study believes that identifying the learners’ preferred learning styles will help educators choose the most effective teaching methods.
  • Thumbnail Image
    ItemOpen Access
    Evaluation of Learning Management Systems for Success Factors.
    (Kabale University, 2024) Muhaise, Hussein; Adeyemi, A. L.; Muteb,I.J.
    Evaluating an information system for success is key. Evaluating the success factors of a Learning Management System (LMS) is essential in the perspective of information systems success in a developing country context. eLearning is vital to the educational system considering its benefits and impacts, particularly in accessing learning from remote areas, suitable for different learners’ categories, and minimal resource utilization in terms of cost and time. Through the literature reviewed on eLearning and Information Systems, the desire to determine the variables that measure the success factors for information systems continues. Existing Information System (IS) success models do not sufficiently evaluate eLearning in developing country, Uganda inclusive as the current IS success models are generic in nature. This study aimed to describe a model of information system success tailored to the eLearning system, Kampala International University as a case study, a Uganda’s context. To address the above objective, a field study was conducted, using a questionnaire to determine factors for information systems success in Uganda, a sample size of 370 respondents were used from a population size of 5500 using the Sloven formula. The respondents comprise of 340 students, 27 lecturers and 4 administrators. The identified success factors are skills & training,infrastructure and management support. Data were cleaned and analyzed using Statistical Package for Social Scientists version 20.0 (SPSS). This study adopted Delone and Mclean’s Information System Success model (2016) and extended it using factors obtained from the field study survey. Analysis was performed to evaluate the model. Results of the study showed that all the independent variables management support, infrastructure, skills and training are positively related to the dependent variable of intention to and use of information system. There exists a strong relationship between the multiple independent variables and the dependent variable. All factors identified has a positive impact in explaining the variation in intention to use and use of the system with r coefficients of 0.343, 0.406 and 0.406 respectively. The results of the study presented a model of success factors for the Learning Management System, eLearning specifically. For future research, this study recommends conducting qualitative studies to delve deeper into the nuanced perceptions and experiences of learners and teachers as well as looking into policies that can promote eLearning especially in developing countries.
  • Thumbnail Image
    ItemOpen Access
    The Taxonomy Mobile Learning Applications in Higher Institutions of Learning in Ugandan Universities: A Case of Kabale University, Uganda.
    (Kabale University, 2024) Muhaise, Hussein; Businge,Phelix Mbabazi; Ssemaluulu, Paul; Muhoza, Gloria
    Since the use of mobile devices outpaces that of laptops and desktop computers today, the usability of these mobile devices is an important consideration. When mobile learning (a new kind of electronic learning) takes shape, bringing an important feature of mobility, the trend expands deeper into teaching and learning. Usability describes the quality characteristics of software product usage; hence, usability testing is a crucial concern in developing companies for the success of product deployment and use. The vast majority of existing usability evaluation approaches were created for desktop software development. As a result, currently, existing models do not specifically address mobile learning, presenting a gap that we aimed to remedy. The study developed a model that estimates usability as a function of aggregated usability influencing factors. To provide a more comprehensive model, the proposed model includes essential features from other accessible models and incorporates the majority of those that assist mobile learning. A mobile learning prototype application was designed, tested, and installed to evaluate the efficiency of the developed model, coupled with a task list for objective research. Using a sophisticated statistical technique, the feedback from the experiment and survey was then utilized to assess and validate the prototype application in terms of high, average, or low usability. The findings act as guides for eLearning-developing businesses to create more relevant mobile learning applications with high levels of usability.