http://dlib.iit.ac.lk/xmlui/handle/123456789/18 2026-04-05T19:04:02Z 2026-04-05T19:04:02Z Paragoda Pathirana, Uthpala http://dlib.iit.ac.lk/xmlui/handle/123456789/2828 2025-07-01T08:55:00Z 2024-01-01T00:00:00Z

Prediction of DDoS Attacks in Cloud Computing Environment using Deep Learning Algorithms Paragoda Pathirana, Uthpala "Cybersecurity has become an ever-evolving challenge, as the digital landscape becomes increasingly intertwined with daily life. Distributed Denial of Service (DDoS) is a type of cyber- attack that threatens the availability of online resources. Organizations without proper protection face an average expense of $200,000 for each DDoS attack irrespective of how often, how long, or how extensive the attacks are. Even small-scale businesses bear a significant impact because of DDoS attacks, incurring an average recovery cost of $120,000. The study explores the landscape of DDoS attacks, sheds light on the evolving nature and the impact of such attacks. Furthermore, by drawing inspiration from remarkable capabilities of deep learning, this research proposes a novel solution for DDoS attack detection. By using the capabilities such as adaptability and pattern recognition prowess of deep learning models, the aim is to improve the accuracy and precision of network intrusion detection systems. In conclusion, the suggested deep learning technique shows promise as a way to strengthen cyber defences and provide a proactive method of detecting DDoS attacks. In the end, the research's findings promote a safer and more secure digital world by adding to the continuing conversation on cybersecurity resilience in the face of new threats."

2024-01-01T00:00:00Z Jayasinghe, Thilina http://dlib.iit.ac.lk/xmlui/handle/123456789/2827 2025-07-01T08:51:16Z 2024-01-01T00:00:00Z

Forecasting For the West Texas Intermediate Crude Oil Prices with Extreme Values Treatment Jayasinghe, Thilina "Crude oil, often called ""oil,"" is a fossil fuel and a naturally occurring liquid hydrocarbon found underground in different geological formations. It has various types of grades; the West Texas Intermediate Index was used in this study. Usually, fluctuations in crude oil prices affect people, and lead governments to manage their budgets and make policies, while businesses and firms can plan for things like production and investments in either a negative or positive way. Therefore, forecasting crude oil prices for the future has become a major topic in the investment field. Hence, this study was conducted to answer those questions. The main objective of this study is to compare the predictive accuracy of the top-performing extreme value detection technique and the top-performing extreme value replacement technique by using classical time series models and deep learning architectures to forecast West Texas Intermediate crude oil prices. Therefore, five extreme value detection techniques were used: Fixed Price Threshold, Standard Deviation Filter on Prices, Moving Window Filter on Prices, Recursive Filter on Prices, and Percentage Price Filter. After identifying spikes in the dataset, those were replaced using four different replacement techniques: mean, median, damping scheme, and threshold scheme. After doing all the data preprocessing steps, 21 datasets were finalized for the modeling phase. To be more fair, statistical and deep learning architectures were used in this study. As a statistical model, Autoregressive Integrated Moving Average was selected because the data did not follow any seasonality or stationarity. After fitting the statistical model, deep learning architectures such as Recurrent Neural Networks, Long Short-Term Memory, and Gated Recurrent Units were used, and the best model was selected by referring to the Mean Squared Error and Root Mean Squared Error values. Finally, it was confirmed that the Long Short-Term Memory model with a Fixed Price Threshold with damping scheme replacement was the best out of 21 models. It obtained a Mean Squared Value of 4.23 and a Root Mean Squared Value of 2.03. A Python Flask application was finally developed by taking the best model that can be used to forecast crude oil prices up to 100 days accurately."

2024-01-01T00:00:00Z Yapa, Thejani http://dlib.iit.ac.lk/xmlui/handle/123456789/2826 2025-07-01T08:32:07Z 2024-01-01T00:00:00Z

Extreme Multi Label Classification for query-to-tag recommendation for food blog platform Yapa, Thejani "This research addresses the intricate domain of culinary, query-to-tag recommendation within the context of food blog platforms, with a specific focus on Extreme Multi-Label Classification (XMLC) techniques. The primary objective is to enhance the accuracy, efficiency, and responsiveness of query-to-tag recommendations, ultimately enriching the user experience. To achieve this, a framework implemented offering a versatile and modular approach to deep extreme multi-label learning. The research delves into critical challenges, such as data scarcity, scalability, and real-time response demands, offering solutions that encompass feature architecture design, sub-linear search structures, and efficient shortlisting mechanisms. The operational objectives encompass data collection, model development, and real-time inference mechanisms, leading to the integration of the XMLC-based recommendation system into a food blog platform prototype. The platform showcases seamless functionality, connecting users with pertinent culinary tags and recommendations. While the research achieves its goals, it acknowledges limitations, including data availability constraints and potential scalability issues as the platform evolves. Future enhancements are envisioned, encompassing data augmentation, user-generated content, personalization, real-time updates, and internationalization. In conclusion, this research introduces a framework and application that significantly advance the field of culinary content recommendation. By addressing challenges and charting a course for future improvements, it not only elevates the user experience but also paves the way for innovative culinary exploration in the digital realm."

2024-01-01T00:00:00Z Thillayampalam, Sujith http://dlib.iit.ac.lk/xmlui/handle/123456789/2825 2025-07-01T08:29:10Z 2024-01-01T00:00:00Z

Morphological Characteristics of Galaxies in Studying Galaxy Evolution via Optimising their Classification Thillayampalam, Sujith "In the beginning, the infinite state of energy released had been the source for the formation of stars, galaxies and other celestial bodies. Ever expanding universe has posed questions on its origin and evolution stages to understand our very existence in this universe and the scientific community is looking for new ways to decode the mysteries in galaxies. As a result multiple sources of modern measurements such as photometric study were employed. Galaxies particularly their physical appearance, spectroscopy, multiwavelength, and their inner arrangement are in the focal point to understand their evolution. Also new studies also reveal inter-merges between them which is commonly termed as “merges” reflected in their shapes and colours. This research takes big data mechanisms in resolving morphological complexities to produce an appropriate classification model, even though considerable interest has been shown in recent years. Astronomy in the modern day is a semi-structured study based on facts and figures. Generally these figures are unusually big numbers. In modern era astronomy, big data is an inclusion as the 4th paradigm in scientific research due to the reasons that large sky surveys are conducted. Researchers have started to look into the facts with survey images, redshift, multiwavelength, temperature measurements which adds more volume on a daily basis to the existing repositories of data. Exponential growth of information in a small period of time has added both complications as well as new opportunities to zoom in for refining scientific study."

2024-01-01T00:00:00Z