Classification of Customer Loans Using Hybrid Data Mining

Jurnal Informatika

Sarjon Defit¹  Eva Rianti¹  Eka Praja Wiyata Mandala¹ 

[1] Universitas Putra Indonesia YPTK Padang;

关键词: clustering;    classification;    customer loans;    hybrid data mining;   

DOI  :  10.30595/juita.v10i1.12521

来源: DOAJ

【 摘 要 】

At this time, loans are one of the products offered by banks to their customers. BPR is an abbreviation of Bank Perkreditan Rakyat. BPR is one of the banks that provide loans to their customers. The problem that occurs is that the number of loans given to customers is often not on target and does not meet the criteria. We propose a hybrid data mining method which consists of two phases, first, we will cluster the eligibility of customers to be given a loan using the k-means algorithm, second, we will classify the loan amount using data from the clustering of eligible customers using k-nearest neighbors. As a result of this study, we were able to cluster 25 customers into 2 clusters, 10 customers into the "Not Feasible" cluster, 15 customers into the "Feasible" cluster. Then we also succeeded in classifying customers who applied for new loans with occupation is Entrepreneur, salary is ≥ IDR 5000000, loan guarantees  Proof of Vehicle Owner, account balance is < IDR 5000000 and family members is ≥ 4. And the results, classified as Loans with a small amount. We obtained the level of validity of the data testing of each input variable to the target variable reached 97.57%.

【 授权许可】

Unknown