By using a refined K-means algorithm that includes optimized initial centroid selection and lessened distance computations, this research clusters students based on their campus behavior patterns utilizing a sample of 324 students who are enrolled at the XX Vocational College. Associations between these behavior patterns and academic results are then computed through an improved version of the Apriori algorithm. To optimize SVM parameters, a fruit fly optimization algorithm (FOA) is presented to allow early detection of students at academic risk. Main observations indicate that most students spend between 600 and 900 yuan per month, with the average being 789.37 yuan. Internet fees on the campus are mostly 37.64 yuan per month (43.52 percent), but it has been seen that the cost ranges between 9 and 48 yuan. Frequency of bathing is 9-17 per month in 48.77 percent of the sample and the lowest book borrowing group is 73.15 percent of the students who borrow an average of only 6.67 books each. Daily living habits and academic engagement were found as the main determinants of academic performance among the behavioral dimensions evaluated, with spending patterns having relatively low predictive power. It is worth noting that irregular routines seem to result in increased expenditure, implying that lifestyle discipline affects financial behavior too. The suggested model shows high fitting precision and low prediction error, providing a consistent model to be used by vocational college administrators to establish a constant loop of monitoring, early warning, specific intervention, and systematic improvement when it comes to student development.
