Downloads: 3 | Views: 53 | Weekly Hits: ⮙3 | Monthly Hits: ⮙3
Study Papers | Software Engineering | Zimbabwe | Volume 12 Issue 11, November 2023
Developing Datasets from Limited or Scarce Data Sources
Tawanda Madzidzise | Monika Gondo [3]
Abstract: Artificial intelligence(AI) is one of the best avenues of business success since there is a lot that the approach can achieve from data analysis, projections, insights and many more, this approach is likely to boost the world economy by $15.7 trillion by 2030, (?AI to Boost World Economy by over 15 Trillion Dollars in Seven Years? n.d.).In a scenario where one identifies a gap or a problem, and will need to embark on the solution in Artificial Intelligence, the very question will there be a dataset(s) that will be used for developing, training and testing the models or solution. This is one of the major limitations in developing solutions in the health sector in Zimbabwe mainly in the subsection of mental health where there is limited datasets which can be used in coming up with models to cater for the problems in mental health which include early detection of depression, anxiety and other problems which are in mental health.This should not be a major issue since we can have solutions which can provide solutions for scenarios like these, the research will identify possible drawbacks in developing models in scenarios of limited/scarce data targeting the case of mental health in Zimbabwe. There is a chance to develop model that speak to Mental Health in Zimbabwe, thus through finding ways to do so in scarce data environment. This will be was the major focus of the research and it was found to be achievable, then the researcher went on to develop a dataset with a limited number of data which was then used to train a model through different algorithms and it proved to be feasible at last. The result was great but there were challenges which were faced and some proposed solutions listed and further research proposals so as to attain better results and improve on the issue of bias.
Keywords: model, AI, Mental health, data
Edition: Volume 12 Issue 11, November 2023,
Pages: 1984 - 1988