Browsing by Author "Perera, V. A. S."

Now showing 1 - 3 of 3

Embargo
A Machine Learning Approach to Predict Default Lease Cases in Sri Lankan Financial Institutions
(IEEE, 2022-12-26) Perera, V. A. S.; Kavirathne, G. P. R. A.; Karunathunge, L. C. R.; Dewapura, B. N.; Karunasena, A.; Pemadasa, M. G. N. M.
The economic growth of a country can be aided by a strong financial services industry. Therefore, financial companies play a vital role in today’s society. However, by providing credit facilities, they expose themselves to a significant amount of risks, since most of them lack a proper strategy to identify whether the customer is reliable and capable of paying back on time. Hence, it is widely acknowledged that having a proper strategy in place to manage and lessen the credit risks that these companies face is more beneficial, rather than relying on traditional manual techniques. This study is intended to propose a machine learning-based solution to predict possible financial lease defaults beforehand. The dataset used in this work was obtained from a leading finance company in Sri Lanka, where the data were related to leasing contracts and their equipment. According to the final results of this study, a deep learning model implemented using an Artificial Neural Network, which was compared against several other machine learning models, is the best to predict default lease cases in Sri Lankan financial institutions. The finalized model provides 93.93% of classification accuracy, 85.49% of F-measure, 87.69% of AUROC score, and 80.41% of Kappa score.
Embargo
A Machine Learning Approach to Predict the Personalized Next Payment Date of An Online Payment Platform
(IEEE, 2022-12-09) Karunathunge, L. C. R.; Dewapura, B. N.; Perera, V. A. S.; Kavirathne, G. P. R. A.; Karunasena, A.; Pemadasa, M. G. N.
Use of digital payments has risen exponentially in the recent past especially due to the COVID-19 pandemic. This is because online payment methods offer many benefits in performing their day-to-day transactions and paying utility bills such as electricity bills, water bills, telephone bills and etc. Knowing when a consumer will perform a specific online transaction, or bill payment is beneficial to an online payment platform to plan marketing campaigns since targeted marketing has become very prevalent nowadays. However, predicting this is not an easy task since thousands of transactions are happening in each and every minute of an online payment platform. This paper presents the results of a study that investigated predicting the customer personalized, utility bill payment type wise next payment date of a financial company in Sri Lanka by using machine learning techniques. This is accomplished by analyzing not only online transaction history but also customer characteristics and a holiday calendar which is specific to Sri Lanka. At the end of the study, it was identified that XGBoost Regressor is the most suitable machine learning algorithm, etc deal with this scenario which provided 91.02% accuracy. These predictions will be used for sending personalized reminders and discount offers to customers without sending general common notifications when they are planning to do an online payment. Such reminders and offers will be notified on the mobile devices of the customers and, ultimately both customers and the business owners will be benefited by this.
Embargo
A Meta-learning approach to Predict Non-performing Loans in Sri Lankan Financial Institutions
(IEEE, 2022-12-26) Kavirathne, G. P. R. A.; Perera, V. A. S.; Karunathunge, L. C. R.; Dewapura, B. N.; Karunasena, A; Pemadasa, M. G. N. M.
Most financial institutions make the majority of their income from loan interest. However, due to the current financial crisis in Sri Lanka, non-performing loans have been the focus of financial industry concerns. Before the crisis, financial institutions were more ready to lend to businesses and individuals. Since the crisis, the rate of non-performing loans has increased, limiting the company’s growth. Predicting the likelihood of nonperforming loans can help in lowering this credit risk. Therefore, this paper presents a machine learning approach to predict nonperforming loans of a financial company in Sri Lanka. Moreover, this study also attempts to develop a meta-model that combines various classifiers, including K-Nearest Neighbors, Decision Tree, Random Forest, Support Vector Machine, and Naïve Bayes. The meta-model is also compared with the baseline models. The predictive performance of all models is compared using accuracy, precision, recall, F1-score, and AUROC score. At the end of the study, it was identified that the meta-learning model is the most effective model to handle this case, with a classification accuracy of 93.09%, precision of 82.68%, recall of 92.17%, F1 score of 86.24%, AUROC score of 96.0%, sensitivity of 92.44% and specificity of 93.17%.