إستخدام خوارزميات تعلم الآلة غير الخاضعة للإشراف لدراسة أساليب التعلم والخوف من الفشل للكشف عن أنماط تكيف طلاب المستوى الثاني بكالوريوس stem في الجامعات المصرية
The aim of the research was to detect clusters of students based on the variables of fear of failure and six learning styles, utilizing cluster analysis as one of the unsupervised machine learning algorithms. The research sample comprised 183 male and female students, with 40 (21.9%) being male and 143 (78.1%) being female. All participants were enrolled in the second level of STEM bachelor's degree programs offered by five Egyptian universities that provide this type of education. The results revealed the presence of two distinct clusters. The first cluster consisted of students who were not adapted to STEM education. They were characterized by a high fear of failure, a lack of preference for participatory learning, low enthusiasm for learning, low motivation, lack of self-confidence, and difficulty meeting learning requirements. The second cluster comprised students who were adapted to STEM education. They exhibited no fear of failure, preferred participatory learning, displayed enthusiasm for learning, had high motivation and self-confidence, and successfully met the required learning requirements. Furthermore, the research found that the prevailing learning style among students was the one adapted to STEM education, accounting for 68.31% of the sample. The relationship between an individual's belongingness to one of the two clusters and certain demographic variables (specialization, gender, university) was also examined. The results concluded that gender and university affiliation were not statistically significant factors associated with belongingness to either cluster. However, the student's scientific specialization showed a statistically significant association with cluster membership. Specifically, students specializing in mathematics were less likely to adapt to STEM education compared to students in other specializations. A case study was conducted to explore the reasons behind this finding. (Published abstract)