Fraud prediction in loan default using support vector machine

Abstract

The concept of taking loan has been in existence since inception of the human race but it is now taking diverse dimensions. This spans through personal exchange of loans for repayment based on personal track records, enjoying loans as proceeds of daily contribution without collaterals, except for the banking sector that requests collaterals for official loans. The uniform occurrence of being unable to pay the debts and resulting in a default is evident to the level of bank closures and nations’ bankruptcy is experienced across board. With a large volume and variety of data, credit history judgment by man is inefficient; case-based, analogy-based reasoning and statistical approaches have been employed but the 21st century fraudulent attempts cannot be discovered by these approaches, hence; the machine learning approach using the support vector machine. This work employs a supervised learning approach based on machine learning to predict the possibility of a fraud in a loan application through hidden trends in data instead of giving loans which ordinarily should not be approved; past occurrences discovered through machine learning reveals risky loans and a possible fraud by humans in approvals that can result in a default. Machine learning approaches are able to detect fraudulent financial statements to avert business comatose.