The empirical research of teamwork competency factors and prediction on academic achievement using machine learning for students in Thailand

Introduction. Teamwork competencies are important for achieving the success of education, developing a meaningful and lifelong career, bringing new ideas, helping to solve problems, academic achievement, and building morale. the current problems with teamwork competency factors are still vague and unclear, and lack of predictive data on education. The research objectives were to analyse factors of students’ teamwork competencies, to confirm factors of students’ teamwork competencies, and to predict the effect of teamwork competencies on academic achievement using machine learning. Methods. Participants were students in an advanced professional innovation programme undergoing an empowerment process. They were divided into two groups: 205 students for exploratory factor analysis achieved through stratified random sampling and 418 students for confirmatory factor analysis, correlation analysis, and prediction of academic achievement, also achieved via stratified random sampling. Instruments included a teamwork competency self-assessment digital form, which had item-total correlations between 0.67 and 0.77 and a Cronbach’s alpha coefficient of reliability equal to 0.98. The data were analysed by summary statistics (mean, standard deviation), correlation analysis, exploratory factor analysis, confirmatory factor analysis, regression analysis and decision trees. Results. The results showed that teamwork competencies factors included building a team relationship (BTR), participation in team exchanges (PTE), adapting and creating a team atmosphere (ACT), and supporting a team (STE), which explained 65.752% of the total variance. Teamwork competency factors fit the empirical data (chi-square=618.54, df=565, p=0.059, GFI=0.93, AGFI=0.90, RMSEA=0.015), and teamwork competency sub-factors were correlated with the total score by 0.97, 0.94, 0.95 and 0.92, respectively. In addition, the machine learning predicted by regression BTR, PTE, and ACT were predictive of academic achievement, explaining 60.80% of the variance in this variable, and predicted by decision trees ACT and PTE were predictive of academic achievement as well. Practical significance. The novelty created by this research is that innovative product of teamwork competencies factor, that is suitable for a studied and modern context, and prediction of machine learning that will increase empirical strength and create a methodological novelty of empirical research that uses multi-stage, multi-methods and robust validation. The utility of practice, the colleges apply to students for accurate measurement, good academic achievement and leads to reflection and improvement of teamwork for performance efficacy, bringing success to individuals and society.