http://dlib.iit.ac.lk/xmlui/handle/123456789/1619 Wed, 15 Apr 2026 06:33:13 GMT 2026-04-15T06:33:13Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2182 Loan Default Prediction in Abc Finance Company a Machine Learning Case Study Palihadaru, Thamali Finance companies equally play a vital role in the Financial System in Sri Lanka. Loans and Advances represent the majority of the asset class of the financial position which accounted for 76.8% of the total assets in 2021. While financing through credit being a driving force of the economy, it has become the biggest risk for any financial institution. Non- Performing Loans (NPL) of the Non- Banking Finance sector have also been on the rise, reaching 13.9% and 11% in 2020 and 2021. As a solution to the problem statement, we have conducted a case study on employing machine learning models for credit risk modeling in a well reputed Finance Company in Sri Lanka. Five machine learning models namely, Logistic Regression, Decision trees, Random Forest, XGboost and Adaboost were used and evaluated the performance of each model. It was revealed that XGBoost model outperforms the other models with the highest model performance. (Accuracy (0.78), Weighted Avg Precision (0.87), Weighted Avg Recall (0.77), Weighted Avg F1-score (0.82) and AUC-ROC (0.60)). It was evident that feature selection and hyperparameter optimization will impact the performance of the model. Correlation coefficient heatmap, Chi- Squared test of independence, Select K best, Recursive Feature Elimination and Random Forest Feature importance were used as feature selection. The highest performances were shown with Random Forest feature importance's. It was observed that the importance of features namely, No of Rental (Term), Effective Rate, Rental, Age, Income, LTV and vehicle age are found to be significantly high. Addition to that, we discuss model interpretability using LIME method. Sun, 01 Jan 2023 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/2182 2023-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2181 Predicting the Turnover of Air Freight Logistics Based on The Economic Factors in Sri Lanka Alwis, Sasanka In today's highly competitive business world, the ability to predict revenue is crucial for decision-making and strategic planning. This research aims to develop a predictive model for revenue rate using machine learning algorithms and economic indicators. The study focuses on a freight transportation company that operates in multiple countries and relies heavily on economic indicators such as inflation, interest rates, exchange rates, and gross domestic product (GDP) to forecast revenue. Data from the International Monetary Fund (IMF) and Trading Economics data library was collected for 10 different economic indicators. A linear regression model was developed to determine the significance of these indicators on revenue rate. The model identified six significant predictors: annual interest rate, GDP annual million, inflation quarter, quarterly interest rate, CPI quarter, and exchange rate. The model achieved a coefficient of determination (R2 ) of 0.39, indicating that the selected economic indicators explain 39% of the variance in revenue rate. To make the predictive model more accurate, three machine learning algorithms were evaluated: linear regression, decision tree, and random forest regression. The models were evaluated using cross-validation techniques, and the linear regression model was selected as the best model, achieving a mean absolute error of 6.82. A dashboard was created using Tableau, which allows the management to input values for selected economic indicators, and then the model predicts the revenue rate for the company. The dashboard also includes links to IMF data and Trading Economics data library, providing additional economic data for further analysis. Overall, this research demonstrates the potential of machine learning algorithms to predict revenue rate using economic indicators. The study contributes to the field of revenue prediction models by using machine learning algorithms, which can produce more accurate and reliable results. Future research could expand on this study by including additional economic indicators or exploring other machine learning algorithms to improve the predictive power of the model Sun, 01 Jan 2023 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/2181 2023-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2180 Predicting Undergraduates Dropouts Using Classification Techniques Perera, Nipuni "Undergraduate dropout is one of the biggest concerns in higher education institutes. This has become a significant concern locally as well as globally. Student retention has gained more attention from university administrators, especially those in private-sector higher education institutes as the competition is quite high in the private sector. This research’s main objective is to predict undergraduate dropouts in the Information Technology Degree Program of a non-state higher education institution in Sri Lanka. Logistic Regression, Random Forest, Naïve Bayes, Artificial Neural Network (ANN), Decision Tree, and Support Vector Machine (SVM) classification techniques were used for the prediction. According to the results, SVM has the best F1 score which is 90%, ANN, Decision Tree, and Logistic regression got 88%, and Random Forest and Naïve Bayes have an 87% of F1 score. It has also been identified that dropouts are high in those who have done Advanced Level in Art Stream and under Other Category. Therefore, before students get register from those categories if faculty can give them an aptitude test and select the relevant candidates, will be helpful to reduce the dropouts. Data mining techniques can improve the quality of education in non-state higher education institutes as this helps to identify the hidden patterns of educationally linked data." Sun, 01 Jan 2023 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/2180 2023-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2179 Classification of Cyberbullying Romanized Sinhala Comments in Online Platforms Hettiarachchi, Yasindi "This study investigates the detection of cyberbullying in Romanized Sinhala using various machine learning classifiers and feature extraction methods. The primary objective is to identify the most effective combination of classifier and feature extraction techniques for this task. We employ rule-based, Bag-of-Words (BoW), and Term Frequency-Inverse Document Frequency (TF-IDF) feature extraction methods, as well as additional features such as word count and gender. The classifiers studied include K-Nearest Neighbours (KNN), Voting, Random Forest, Support Vector Machines (SVM), Decision Tree, Naive Bayes, Multilayer Perceptron (MLP), AdaBoost, and Logistic Regression." Sun, 01 Jan 2023 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/2179 2023-01-01T00:00:00Z