Open AccessDATA MINING PADA PENENTUAN KELAYAKAN KREDIT MENGGUNAKAN ALGORITMA K-NN BERBASIS FORWARD SELECTION DATA MINING ON CREDIT FEASIBILITY DETERMINATION USING K-NN ALGORITHM BASED ON FORWARD SELECTION
Author(s) -
Sitti Harlina
Publication year - 2018
Publication title -
creative communication and innovative technology journal
Language(s) - English
Resource type - Journals
ISSN - 2655-4275
DOI - 10.33050/ccit.v11i2.591
Subject(s) - selection (genetic algorithm) , feature selection , credit risk , credit rating , computer science , finance , algorithm , credit score , value (mathematics) , data mining , actuarial science , business , machine learning
Along with the development of economic and banking services, the term credit is one of the banking products, in determining credit worthiness required data, information about creditworthiness is an interesting issue to be in carefully. In line with business growth, the leasing or crediting sector (credit), has complex and complex problems and problems in which the data are grouped into two classes, good credit and bad credit. In banking activities the feasibility of credit determination is often the problem of bad debts caused by the failure of partial return of loans provided to the borrowers for it required accurate data prior to lending or credit. So it is appropriate to use data mining classification techniques. This study discusses the Data Mining on the determination of creditworthiness using the K-nn-based Forward Selection algorithm. K-nn algorithm applied to consumer data using credit financial services with the help of Forward Selection. The experimental results show that K-nn in classifying credit risk has accuracy value for K-nn at K = 11 accuracy = 68,30%, and for Forward Selection at K = 9 which accuracy level = 73,60%, using feature Forward Selection can produce a good degree of accuracy, so it can be applied to determine the risk of credit risk eligibility.