http://dlib.iit.ac.lk/xmlui/handle/123456789/2801 2026-04-08T18:29:05Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2860 Predicting Peripheral Sensory Neuropathy Among Patients With Type 2 Diabetes Mellitus in Ratmalana Gamage, Yasangi "Peripheral Sensory Neuropathy (PSN) is a serious complication of Type 2 Diabetes Mellitus (T2DM) that can lead to severe health issues, including foot ulceration and amputation. This research aims to develop a predictive model to identify risk factors for PSN among T2DM patients in Ratmalana, Sri Lanka, using data collected from the ABC Diabetes Clinic. The study will involve 350 T2DM patients and will analyze variables such as age, gender, smoking history, alcohol consumption, blood sugar control (HbA1c levels), family history of diabetes, and diabetes duration. Advanced machine learning techniques and data analysis will be employed to detect early signs of PSN and provide insights for healthcare interventions. The methodology includes data preprocessing, feature selection using the Chi-square method, and model evaluation through metrics like Area Under the Receiver Operating Characteristic Curve (AUC) and F1-Score. The research is structured into eight chapters, starting with an introduction to PSN's significance in T2DM and the research objectives. A literature review follows, synthesizing existing knowledge and identifying gaps. Methodology details data collection, preprocessing, and model development. Subsequent chapters focus on feature engineering, machine learning model selection, training/testing phases, and ethical considerations in diabetes prediction research. The findings are expected to enhance early PSN detection and risk factor identification among T2DM patients, enabling timely interventions that improve quality of life. This research will also provide a framework for expansion to other healthcare facilities and regions, contributing significantly to diabetes management and policy-making efforts aimed at reducing the burden of PSN in similar contexts" 2024-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2859 Predicting the Financial Distress of Licensed Finance Companies Operating in Sri Lanka using Machine Learning Techniques Karunathilaka, Withanage Dinithi Nimesha "This study investigates the application of Machine Learning techniques in predicting the financial distress of Licensed Finance Companies (LFCs) operating in Sri Lanka. A variety of Machine Learning models including Logistic Regression, Decision Tree, Support Vector Machine, Naïve Bayes, XGBoost, Neural Networks, and Random Forest were evaluated for their predictive power, on data pertaining to key financial soundness indicators of LFCs categorized under the CAMEL Framework spanning over the past 13 years from 2010-2022 on a quarterly basis. The study demonstrated that advanced models such as XGBoost and Random Forests outperformed the simple and traditional techniques in terms of accuracy and predictive power. The analysis included a feature importance assessment to provide insight into the key financial ratios contributing to financial distress prediction and aiding in providing actionable recommendations for the regulatory entities based on the prediction outcomes of the best-performing models. Furthermore, the study underscores the importance of timely and accurate predictions in the financial sector, as early detection of distress can help mitigate systemic risks. The findings suggest that incorporating ML tools can enhance regulatory oversight by offering precise, data-driven recommendations for policymakers. Despite the promising results, the study acknowledged several limitations regarding data availability and model interpretability. This research contributes to the literature on financial risk management by highlighting the potential of integrating Machine Learning techniques in the financial domain to promptly detect financial distress among the LFC sector paving the way to avoid potential damage to the financial system and ultimately to the public. " 2024-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2858 Supplier Risk Prediction Model Weerasinghe, Vihanga "Supply chain networks are the backbone of the majority of businesses globally. Due to the importance of supply chain networks, the entire global trading community is focused on ensuring supply chain resilience and sustainability. However, there are multiple supply chain risks that constantly disrupt the smooth functioning of supply chains. Supplier-related risks are one such major type of risk which could be controlled by organizations in a much more effective manner, rather than other types of risks, due to the level of control the company holds in the supplier selection process. Thus, this study has drawn its focus to develop a Supplier Risk Prediction model, which will benefit the organizations in accurately identifying the levels of supplier risk in ‘High’, ‘Medium’ or ‘Low’ categories. This model is recommended to be deployed in a strategic position between the sourcing process and the PO issuance stage to ensure that the businesses select the most optimal suppliers. The main objective of this study is to develop a Supplier Risk Prediction model to accurately predict the supplier risk levels. In order to execute the study a secondary dataset was selected to aid the development of the model. R Studio is used as the data analysis tool for this quantitative study. Initially the dataset was preprocessed, then the model was trained and tested. Five main classification techniques namely, Random Forest, SVM, Decision Tree, Logistic Regression and kNN technique were used for model development. Based on the measures model comparison, SVM emerged as the most compatible and predictive model with the highest accuracy in predicting the supplier risk levels. Afterwards, the SVM Model was applied on an actual dataset to validate the model." 2024-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2857 Developing a Predictive Model to Forecast the Future Financial Performance of Sri Lankan Companies in the ‘materials’ Industry. Liyanage, Upeksha "This research focuses on developing a predictive model to forecast future stock prices of companies within the ""Materials"" industry listed in the “Colombo Stock Exchange”, by leveraging financial ratios. Utilizing a robust dataset comprising various financial ratios—such as Price-to-Earnings Ratio, Debt-to-Equity Ratio, Return on Equity, and many others—this study aims to identify the most significant predictors of stock price movements within this sector. By employing advanced statistical techniques and machine learning algorithms through tools such as Pycharm for coding and Power BI for data visualization and analysis, we systematically analyze historical financial data to construct a model that can accurately predict future stock prices. The study begins with a comprehensive literature review to identify previously established correlations between financial ratios and stock performance. Following data collection and preprocessing, we apply multiple regression analysis & decision tree regression to evaluate the predictive power of each financial ratio. The model's performance is assessed using a split-sample test, with a focus on metrics such as R-squared, mean squared error (MSE), and accuracy percentage to ensure reliability and validity. Our findings reveal that certain financial ratios hold significant predictive capability for stock prices in the Materials industry, offering insights into the financial health and operational efficiency of firms within this sector. The predictive model developed in this research not only enhances investment decision-making but also contributes to the academic literature by providing a focused analysis on the Materials industry. Moreover, it offers a framework that can be adapted and applied to other sectors for forecasting stock prices based on financial health indicators. Through this study, we demonstrate the practical applications of financial ratios in stock market analysis and the potential for predictive analytics in enhancing market efficiency and investment strategies." 2024-01-01T00:00:00Z