Evaluation of a machine learning-based career guidance information system on enhancing career guidance practices in Secondary Schools.
- Author
- Rongoti, Isaac T.
- Title
- Evaluation of a machine learning-based career guidance information system on enhancing career guidance practices in Secondary Schools.
- Abstract
- Access to structured career counselling remains a significant issue for many secondary school students in Zimbabwe, particularly in less-privileged areas. Traditional guidance methods like career days and one-on-one counselling are mostly unavailable due to financial constraints, poor infrastructure, and a lack of trained counsellors. In order to bridge this gap, this study proposes and develops a machine learning-based mobile career guidance information system that provides personalised career path forecasts and mentor recommendations. The system uses a random forest classification model, which was trained on student profiles that include interests, hobbies, and subject strengths, to predict suitable career fields. Additionally, based on profile similarities, a cosine similarity-based content-based filtering method pairs students with suitable career mentors. The system's front end was constructed with Flutter, and the backend authentication was handled by Supabase. A research science methodology was used for this study since the research involved developing, designing and evaluating a technological solution. A purposive random sampling technique was used to select respondents in Bindura District of Zimbabwe, and a mixed-methods approach was the research design that was employed. Structured questionnaires were used to collect quantitative data in order to evaluate students' access to career counselling and validate prediction results. Descriptive statistics were used to gain a comprehensive evaluation of students responds. To assess the system's usefulness and relevance, qualitative data was gathered from open-ended comments provided by teachers and system users. Metrics like accuracy, precision, recall, F1-score and confusion matrix were used to evaluate model performance for system evaluation, while Supabase logs and in-app user interactions enabled formative assessment of system engagement. The proposed system can significantly expand students' access to career counselling in remote and underserved areas, according to the results. It also shows promise in providing customised recommendations that enhance students' academic preferences and abilities. The results of the study show that when machine learning technologies are mobile-accessible and locally contextualized, they offer a feasible means of enhancing career counseling practices in low-resource settings like Zimbabwe.
- Date
- June 2025
- Publisher
- BUSE
- Keywords
- Evaluation
- Career Guidance Practices
- Career Guidance Information System
- Supervisor
- Mr. C. Zano
- Item sets
- Department of Computer Science
- Media
-
Rongoti, Isaac T..pdf