Open Journal Systems

AdaBoost-SVM and Feature Selection of Genetic Algorithm Combination to Enhance Indonesian P2P Lending Credit Risk Assessment

Devon Ho, Ira Fachira

Abstract


Abstract. Many Fintech start-ups were established over the past few years in Indonesia. They saw the opportunity that SMEs needed them as platform to provide loans and fundings in order to run their businesses. However, in that sense, many Fintech companies were considered failed to payback its lenders because they could not gain a sufficient amount of borrowers to join the P2P lending platform. As a result, they experienced an increase in loan loss rate, as well as non-performing loans ratio. That was the reason why many Fintech could not survive and were forced to close their business. Well-performing credit risk management is one of the work that Fintech companies should do. Machine learnings have already been utilized to enhance credit risk assessment, although they still need improvement following the development and changes towards modernization. Some of the techniques such as AdaBoost-SVM and Genetic Algorithm will be discussed in this paper and the author will analyse the effectiveness of both techniques and discover the possibility of having them combined to generate the best outcome.

Keywords: Fintech; non-performing loans ratio; AdaBoost-SVM; Genetic Algorithm;credit risk assessment; machine learnings.

 


Full Text:

PDF

Refbacks

  • There are currently no refbacks.