Enhancing Direct Marketing and Loan Application Assessment Using Data Mining: A Case Study of Yaa Asantewaa Rural Bank Limited

Loading...
Thumbnail Image
Date
April, 2016
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Due to the wide availability of the web and the internet, ever increasing processing power and the continuous decline in the cost of storage devices, data are being generated massively today than it were decades ago. With fields like the banking industry, data are being generated massively on regular basis. So Managers and Administrators are finding ways to turn these data into very beneficial information to their advantage. Marketing risk and Credit risk are considered the most serious problems of every bank. Fortunately all marketing promotions are highly dependent on the data about customers stored in electronic format. Past events contain patterns that are hidden in the data that records them. Data mining, machine learning and artificial intelligence promises the use of models to go through and analyze huge data that are very difficult for human experts. Data mining also possesses incredible power to detect hidden relationships, correlations and associations in data. This study was conducted to demonstrate with practical methods, experiments and datasets that data mining can be used to assist in direct marketing and analyze credit risk assessment. The two algorithms used were decision tree and random forest. The University of Waikato data mining open source software (Weka) was used to simulate all the experiments. Confusion matrix was used to calculate the accuracy, sensitivity and specificity which were used to evaluate the performance of the models. Receiver Operating Characteristic curve was also used to pictorially display the performance of the models. The results of the experiment showed with high precision that the models can be used in detecting prospects for marketing campaigns and also detecting risky loan applicants. The decision tree classifier produced an accuracy of 92.5% and the random forest classifier produced an accuracy of 76.4 %.
Description
A Thesis submitted to the Department of Computer Science, College of Science, Kwame Nkrumah University of Science and Technology in partial fulfillment of the requirements for the degree of Master of Philosophy,
Keywords
Citation