Systematic Literature Review on Pedagogies and Visualization Tools for Machine Learning in K-12 Schools

Abstract

Though studies have been done on Machine  Learning, almost all the studies focused on higher educational institutions, with little attention to K-12 educational settings. Those studies that focused on K-12 are scattered, making it difficult to specifically know which visualization tools best enhance Machine Learning in K-12 schools. This study, therefore, through a systematic literature review determines which visualization tools best promote Machine Learning in K-12 schools. The study specifically considered, barriers to the use of Machine Learning in K-12 schools,  visualization tools for Machine Learning in K-12 schools, and pedagogical strategies that benefit the teaching and learning of Machine Learning in K-12 schools. The study sourced articles from Scopus and the Web of Science database after applying the inclusion and exclusion criteria. Data from the articles were extracted based on the PICO framework and their quality was assessed using the Critical Appraisal Skills Programme (CASP) model. The barriers to Machine Learning in K-12 schools include a lack of information about the development and usage of the tools, selection, and coordination barriers, lack of attention to machine learning by educational stakeholders, and programming demands.  Appropriate visualization tools for Machine Learning in K-12 schools include  MLflow and NN-SVG. Though there exist numerous approaches for teaching ML in K-12 settings such as active learning, inquiry-based, participatory learning, and design-oriented approaches, the best pedagogy that supports machine learning in K-12 schools as per existing literature is participatory learning. Teachers need to acquire the appropriate and specific information and technical know-how or skills about machine learning for promoting visualization lessons in K-12 schools. All teachers should be sensitized to adopt participatory learning pedagogy to enhance the effective use of machine learning in K-12 schools.  Machine learning should be integrated into teaching and learning in K-12 schools since it is ideal for visualization and experimentation, which are inevitable for effective teaching and learning in K-12 schools.