Javascript must be enabled for the correct page display

Predicting Financial Problems in Arrears Customers using Random Forests

Krieken, D.R.J. van (2014) Predicting Financial Problems in Arrears Customers using Random Forests. Bachelor's Thesis, Artificial Intelligence.

bsc_thesis_drj_van_krieken_201_1.pdf - Published Version

Download (1MB) | Preview
[img] Text
akkoord_KriekenDRJvan.pdf - Other
Restricted to Staff only

Download (99kB)


Many banks have felt the effects of the financial crisis through an increase in the debt of their customers. For the bank and customer it is surely preferred to reverse back into a positive balance as quick and long lasting as possible. This present research will be an attempt in determining predictability of which customers in arrears will have more difficulty in paying back their debts to the bank compared to others. This predictability will support the development of new approaching methods to these customers that are different than other clients. In this thesis we will focus to recognize customers with relevant financial problems. For this research a case study was done at one of the biggest Dutch banks. A sample group of customers whose accounts were liquidated was taken. By using random forests and different data techniques to cope with the imbalanced data set, features were found which allowed prediction of these customers in arrears. We found, among others, that the change in amount of transactions is a feature that predicts financial problems. These features will now be implemented so customers with these flags are identified earlier and will receive more appropriate treatment.

Item Type: Thesis (Bachelor's Thesis)
Degree programme: Artificial Intelligence
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 15 Feb 2018 07:58
Last Modified: 15 Feb 2018 07:58

Actions (login required)

View Item View Item