Institutional-Repository, University of Moratuwa
Welcome to the University of Moratuwa Digital Repository, which houses postgraduate theses and dissertations, research articles presented at conferences by faculties and departments, university-published journal articles and research publications authored by academic staff. This online repository stores, preserves and distributes the University's scholarly work. This service allows University members to share their research with a larger audience.
 Research Publications |  Thesis & Dissertation |  E- Books |
 UoM Journal Publications |  UoM Conference Proceedings |  Articles published in Scimago's Q1 journals |  UoM Research Reports |  Other Articles authored by UoM staff |
Recent Submissions
item: Thesis-Full-text
Neuro symbolic AI for assessing employee mental health
(2025) Wickramasinghe, JADL; Ambegoda, ALATDT
In the rapidly evolving corporate landscape, employee mental well-being has become integral to productivity and organizational success. This thesis introduces a groundbreaking Neuro-Symbolic Artificial Intelligence (NSAI) framework that integrates conversational data analysis to monitor and enhance workplace mental health. At its core, the Mentalisys Health Application leverages H2OWave to provide user-friendly dashboards equipped with real-time sentiment analysis, stress, and depression detection capabilities. A novel Commonsense-Driven Symbolic ReAct-NLI (CSR-NLI) technique, based on OpenAI’s language models, combines symbolic reasoning and natural language inference to uncover causality in workplace communication. Through interactive admin and user-specific dashboards, the system fosters proactive mental health interventions and personalized support, promoting a healthier workplace environment. The study’s primary contribution lies in advancing NSAI for robust causal understanding, going beyond conventional sentiment analysis. Results demonstrate significant potential in improving employee well-being and productivity via timely interventions and precise health risk assessments. This work underscores the transformative role of AI in addressing real-world mental health challenges, driving organizational growth, and enhancing employee satisfaction, while setting a new benchmark for AIdriven solutions in corporate mental health management.
item: Thesis-Full-text
Hybrid deep learning for forex prediction : integrating social media sentiment and technical indicators
(2025) Sampath, MKT; Ambegoda, TD
This study proposed a new way to predict currency prices in the Forex market by combining statistical indicators with social media sentiment analysis. The research classifies tweets from influential financial analysts as positive, negative, or neutral and combines sentiment analysis results with technical indicators derived from historical price behavior. This mix of methods improves pattern recognition in the market and leads to smarter trade choices that are more likely to be right.
The VADER Sentiment library, a powerful sentiment analysis tool, has been employed to analyze Forex-related tweets of the most influential market players. The study identifies the most influential tweets based on impact engagement measures (likes, replies, and retweets) and selects the most relevant technical indicators. A dataset comprising ten technical indicators, along with news tweets from influential players in USD and EURO trades, has been collected for the study.
A prediction model was built using several deep learning approaches including Gated Recurrent Units (GRU), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) networks, Recurrent Neural Networks (RNN), and some combined models. The LSTM-CNN-Attention model, which is a recent and specialized architecture for Forex data, was also considered in our experiment. In addition, the best-performing models were tested both with and without impact ratio - Sentiment Score of Social media Text.
To see which worked best, their performance is compared using key points like Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and the R-squared score. Comparative analysis reflected the weaknesses and strengths of individual models, revealing the effectiveness of the ensemble strategy.
The research also points out its limitations. It includes a detailed look at what other popular models are being used in predicting the Forex market. The results show that combining sentiment analysis with technical indicators improves how accurately it can be predicted Forex prices in this research. The hybrid GRU-LSTM model was the best among the deep learning options in this research experiment. It outperformed all of the earlier methods used by other researchers on the EUR/USD Forex data
item: Conference-Extended-Abstract
A Study on the shift of tourist demand towards low-cost carriers
(Engineering Research Unit, 2025) Jayasingha, A; Jayawardhana, B; Fernando, A
The aviation industry plays a vital role in connecting nations, promoting global trade, and driving the growth of tourism. Over time, the structure of the airline industry has been divided into two main business models, Full service carriers (FSCs) and Low cost carriers (LCCs), each catering to different market segments and passenger expectations. Full service carriers operate on a traditional model that emphasises comfort, quality service, and extensive connectivity providing a range of in-flight amenities. In contrast, Low-Cost Carriers adopt a simplified, cost-efficient operational approach aimed at minimising expenses and offering significantly lower ticket prices. This approach allows them to attract price-sensitive travellers and first-time flyers.
item: Conference-Extended-Abstract
Evaluation of key focus measures for smear analysis
(Engineering Research Unit, 2025) Hewavitharana, DC; Jayathilaka, WADM; Dassanayake, VPC; Amarasinghe, YWR
Reliable focus assessment is crucial in microscopic smear analysis, as automated systems' diagnostic accuracy is heavily reliant on image sharpness. Despite the availability of numerous focus measures (FMs), their applicability across biological domains remains unknown. This study employs a weighted, raw value-based generalization approach to evaluate fifty single focus measures across five publicly available Z-stack microscopy datasets. Ten quantitative criteria were used to assess focus curve fidelity and resilience. The raw-value-based generalization score combines weighted mean performance and inter-dataset stability to provide a complete measure of cross-domain reliability. Many FMs have been proposed, including gradient, statistical-, texture-, and transform-based operators. However, their efficacy varies greatly depending on specimen type, staining process, and optical arrangement. Prior comparative studies [4], [5] typically focused on one or two datasets or imaging modalities, limiting the extent to which their recommendations may be applied to larger smearanalysis contexts.
Recent research has also offered deep learning-based focusing algorithms for digital pathology and whole-slide imaging [5]. While effective, these systems typically require task-specific training data, significant computational resources, and function as black box predictors rather than interpretable attention scores. In contrast, traditional handcrafted FMs remain appealing for incorporation with resource-constrained or older autofocus systems that require transparency, minimal latency, and ease of deployment. Our study presents a comprehensive evaluation of fifty
single-focus measures from five publicly available z-stack microscopy datasets that include haematological and microbiological smears. A raw-value-based generalization methodology is suggested to measure both performance and inter-dataset stability, with the goal of discovering a limited set of FMs that are consistently dependable across a wide range of smear types and imaging conditions. The collected findings assist researchers and developers in identifying the
best focus operators for future biomedical imaging applications.
item: Conference-Extended-Abstract
Analysis of airside and terminal side congestion at Bandaranayaka International Airport
(Engineering Research Unit, 2025) Weerasinghe, D; Jayawardhana, B; Fernando, A
Air transport is vital for global connectivity and economic growth. However, rising demand and limited infrastructure have caused increasing congestion at airports worldwide. Bandaranayaka International Airport (BIA), Sri Lanka’s main gateway, was designed for 6 million passengers annually, but this limit has been exceeded. Although a new terminal is under construction, delays have worsened airside and terminal congestion, creating a major challenge for Sri Lanka’s aviation sector.