Prediction of Student Decisions in Choosing the Type of Bank Using Support Vector Machine (SVM)

Abstract

A bank is an intermediate financial institution authorized to take deposits, lend money, and issue promissory notes or banknotes. In the present day, every adult must have at least one bank account. Additionally, bank services range from regular and hajj savings to large-scale loans. Students, one of the bank’s customers, usually utilize services confined to savings to preserve pocket money received from their parents and ordinary transactions like transfers and payments. Several factors, including the atmosphere, administrative fees, and the accessibility of ATMs and bank branch offices, impact students’ decisions about where to save money. It prevents the bank from predicting which services must be enhanced to encourage customers, particularly students, to select the bank. Therefore, prediction is required to ascertain the students’ choice of bank. This study employed data mining and the Support Vector Machine (LibSVM) algorithm. The quantity of data impacted the outcomes of the SVM classification. In addition, kernel types, k-fold values, and sampling techniques also influenced classification accuracy. LibSVM with a kernel type of RBF, a k-fold of 8, and shuffled sampling classified 200 data with an accuracy of 68.40%.