http://dlib.iit.ac.lk/xmlui/handle/123456789/1061 Wed, 15 Apr 2026 06:33:46 GMT 2026-04-15T06:33:46Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1087 A Research on Capital Structure in Manufacturing Firms in Sri Lanka and Recommendations Based on Pecking Order Theory Wijesekera, Gunawardena Gurusinghe Arachchige Tiran Dileepa Gurusinghe " This study empirically examines whether the pecking order model is able to explain the financing behavior of listed manufacturing companies in the CSE. The research examines the problem of does the pecking order hypothesis explain the financing behavior of listed manufacturing companies in the Colombo Stock Exchange (CSE)? If not what are the recommendations and forecasts we can predict as to how the capital structure should be in future. The sample used in this study consists of 32 manufacturing firms listed in CSE. All the collected data are secondary in nature and they were collected from the firms’ financial statements from financial year 2015/2016 to financial year 2019/2020. A time series analysis is also performed using key variables of 29 companies from 2010/2011 to 2019/2020. The results of the analysis indicate that profitability is the most important variable that explains the changes in debt level. Overall the results of the analysis showed no support for pecking order in Sri Lankan manufacturing sector. The findings of the study indicated that significant part of the variation in debt level can be explained by the tangibility, firm size, profitability and financing deficit. Many Manufacturing forms utilized these low interest to fund their operations in the middle of the pandemic situation in the country. Overall the results of the analysis showed no support for pecking order in Sri Lankan manufacturing sector. Through the findings this study suggests that as this research did not use any specific country factors, it restricted the generalization of the research findings to other developing countries. Therefore, it is recommended to conduct further studies in this area. " Fri, 01 Jan 2021 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/1087 2021-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1086 How to predict the profitability for listed finance public limited companies in Sri Lanka Nicholas, Nishanth " In a developing country like Sri Lanka, it is easy to get affected by global crises. Even in recent months, Sri Lanka faced a forex crisis since the prices of most essential goods have been skyrocketing due to the falling of local currency and high global market prices by the Covid19 pandemic. Not only the directly affected industries, but even the finance company’s leasing percentage have also dropped compared to previous years where the profitability ratios have gone down. In addition, there are impacts to the rates such as inflation rates, exchanges rates, etc. Due to these reasons, companies started to predict their profitability and initiated to create strategic decisions plans accordingly. Half of the finance companies in Sri Lanka suffered a profitability shock. Therefore predicting the profitability ratios became one of the significant aspects, since it will help to set goals and plan, helps to budget, and would help to anticipate change within the market. Also since the inflation rates affect business investment and interest rates in finance public limited companies, it can reduce the value of returns, for research purposes the inflation-adjusted profitability ratios are considered to bring more accuracy. The inflation changed return is the proportion of return that takes into account the time's expansion rate. The reason for the expansion-changed return metric is to uncover the profit from speculation after eliminating the impacts of swelling. Eliminating the impacts of expansion from the arrival of speculation permits the financial backer to see the genuine acquiring capability of the security without outer monetary powers. The inflation-adjusted return is otherwise called the genuine pace of return. In addition, in the conclusion section, different types of models were compared and identified to find the fittest model by using the Root Mean Squared Error and Mean Absolute Percentage Error evaluation methods. This will be done using R studio since R permits rehearsing a wide assortment of measurable and graphical strategies like direct and nonlinear modeling, time-series analysis, classification, classical present tests, clustering, and so forth. R is exceptionally extensible and simple to learn the language and cultivates an environment for measurable predictability and illustrations. On Conclusion, the better model out of selected models was derived using the evaluation methods for each company selected and for each hypothesis. The research was prepared with the intention of providing the listed public finance companies in Sri Lanka with a competitive advantage when making decisions based on the predicted profitability ratios" Fri, 01 Jan 2021 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/1086 2021-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1085 Machine learning and statistical approaches for asthma disease prediction among adult in Sri Lanka Gunawardana, Jayakodi Ralage Nishani Amalka " Background: Asthma is an airway-induced inflammatory lung disease that causes breathlessness, wheezing and regular potentially fatal attacks. Diagnosis of asthma at affordable costs is often challenging due to its variability. A machine learning approach and an appropriate application to provide asthma status prediction would be valuable in clinical practice and self-detection of the disease. Methods: This study utilized data of 596 asthmatics and 5898 non-asthmatics who participated in the Sri Lanka Health and Ageing Survey (SLHAS) during the 2018-2019 period in Sri Lanka. Both doctor-diagnosed asthma and patient-reported asthma were considered when deciding the asthma status of a patient. In this research, thirteen machine learning classification algorithms were built on under-sampled data, and ten algorithms were created using imbalanced data. These include machine learning models such as; Logistic Regression, Support Vector Machine (SVM), Decision Tree, Random Forest, Naïve Bayes, K-Nearest Neighbors (KNN), Gradient Boost, XGBoost, AdaBoost, CatBoost, LightGBM, Multi-Layer Perceptron (MLP), and Probabilistic Neural Network (PNN). The performances of these algorithms were evaluated by employing various measures, including Area Under Curve of Receiver Operating Characteristics (AUC ROC) and confusion matrix related indices. Results: The model comparison showed that a Hybrid version of Logistic Regression and LightGBM obtained the highest model performance with AUC and sensitivity of 0.9062 and 79.85%. The developed Hybrid models take wheeze related parameters, Shortness of Breath (SOB) attacks, coughing attacks, tightness in the chest, nasal allergies, physical activeness, exposure to passive smoking, ethnicity and sector, as input parameters and predicts the asthma status. The web application developed eases the burden of users by allowing them to get their own estimates upon entering data. Conclusion: A combination of Logistic Regression and LightGBM models can be utilized to predict the presence of adult asthma successfully. The proposed expert system helps patients in their diagnosis of asthma in both self-diagnosis and clinical diagnosis aspects." Fri, 01 Jan 2021 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/1085 2021-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1084 Integration of business intelligence into E-commerce industry to forecast the sales Muttiah, Jonathan Vithurshan " E-commerce has grown exponentially in the last two years and the pandemic is one of the important reasons for the huge growth. Moreover, E-commerce has caused tremendous changes in the transformation of the shopping experience from traditional shopping to digital shopping. This was made possible due to the wide range of technologies and computer networks that connect the countries and continents and synchronize them into a global village. Accurate sales prediction plays an essential role in reducing costs and preventing the merchants from holding huge amounts of cash flow due to the waste of resources, replenishment of stocks and loss of profits caused by goods in short supply. This research study attempts to predict future sales for an E-commerce business based on data of historic sales. Firstly, this research carries out a thorough critical analysis of multiple previous research studies to allow the reader to understand the background and the process of this research. Additionally, it analyses the factors impacting sales forecasting and also, uses the knowledge gained from this and proposes four different machine learning models which can be used to forecast the sales based on the historic sales data. The models which are used here are boosted decision tree regression, neural network regression, decision forest regression and bayesian linear regression. Finally, the research paper analyzes the accuracy measures for the above mentioned machine learning models on the available sales data. This study and the results obtained from this research can help e-commerce businesses to have better management in their future operations. " Fri, 01 Jan 2021 00:00:00 GMT http://dlib.iit.ac.lk/xmlui/handle/123456789/1084 2021-01-01T00:00:00Z