Master of Philosophy (M.Phil.)
Permanent URI for this collectionhttp://192.248.9.226/handle/123/18721
Browse
Recent Submissions
- item: Thesis-AbstractHigh-performance multimodal approach for defect identification in knitted and woven fabric(2022) Pallemulla PSH; De Silva C; Sooriyarachchi SFabric inspection is a key quality assurance process in the garment industry as it involves the detection of defects in a fabric roll prior to being sent for production. Many studies have been conducted on defect identification in either knitted or woven fabrics, but only a few have considered both types. In this paper, a method for detecting defects in both knitted and woven fabrics is proposed. The method involves extracting co-occurrence, wavelet and local entropy features from a fabric image and classifying the image as defective or defect-free using a classifier with these features given as input. Five commonly-used classifiers were tested. This method was applied to a dataset with seventeen different types of defects and an overall classification accuracy of 93.31% was achieved by the k-nearest neighbours classifier.
- item: Thesis-AbstractInvestigation and development of fuzzy logic based analytics for data warehousing(2021) Asanka PPGD; Perera ASData warehouse is a widely used technology that provides the employees who take strategic decisions within an enterprise with access to any level of required data. Historically, data warehouses were built on crisp values with a key assumption that one attribute value falls into one nominal value. Fuzzy Logic can be built into the data warehouse by treating the dimension value as weightages of different labels. However, in most of the attempts to implement a fuzzy data warehouse, they were limited and non-comprehensive in the implementation when considering end to end aspects of the data warehouse. Using fuzzy techniques, it is possible to represent fuzzy conceptual information in the original domain, that would lead to better analysis. In this research, different types of fuzzy membership functions are defined using different techniques and data warehouse facts and dimensions are designed accordingly. There can be multiple fuzzy functions for one dimension as well as for one fact table depending on the business domain. Apart from defining fuzzy membership function using data-driven methods, there are other approaches of defining fuzzy membership functions such as a derived method where multiple fuzzy memberships are combined to define several fuzzy membership functions. In the literature reviewed, concepts like ETL and OLAP cube were found to be discussed in a limited manner. Non-function techniques are also identified and addressed in the means of validation, configuration, performance, security, scalability in order to make better usability of the fuzzy data warehouse. The scope of this research revolves around end-to-end features of fuzzy data warehousing starting from data extraction and transformation to data warehouse modeling. Implementing a fuzzy data warehouse, helps to enable users with better analyses. To verify whether the proposed fuzzy data warehouse can be applied, a feasibility study is carried out for the domains in which fuzzy data warehousing can be implemented. Concepts related to the outcome from this research are verified with the use of a Sri Lankan plantation data set for four years. The results show that concepts introduced by this research can be implemented in realistic scenarios.
- item: Thesis-Full-textTraceability management in a devops environment with continuous integration(2019) Rubasinghe ID; Meedeniya DA; Perera GIUSSoftware artefacts traceability is an important factor during the process of software development to analyse changes occur in software components. Traceability improves the quality attributes of software systems such that strengthens the testability, maintainability, reusability and helps for the system acceptance by providing consistent system documentation to the users. Meanwhile, the concept DevOps motivates towards the reduction of the gap between development and operations requiring considerable organizational changes. In a DevOps environment, significant software artefact changes are expectable rapidly where continuous integration is essential. Continuous integration is a cornerstone practice in DevOps that frequently merges developer working copies into a single shared branch. There is a requirement of determining and analysing the resulted impact of the traceability in order to make accurate change acceptance decisions during software development. Therefore, the core research problem addressed is determining a methodology for change detection and impact analysis together with software artefact synchronization to preserve consistency across all artefacts in a DevOps environment. A rule-based methodology is followed with visualization and analysis techniques applied on a proof-of-work traceability management prototype tool: SAT-Analyser 2.0. The evaluation results and industry-level user study results have shown the significant usefulness and suitability of the approach to a DevOps environment as well as to any software development process model.
- item: Thesis-Full-textMultimodal user interaction framework for context aware e-commerce(2019) Hewawalpita SGS; Perera GIUSE-commerce has grown up to be a major use of e-services and online purchases through the e-commerce are largely preferred over the traditional brick and mortar purchasing. Yet it is challenging for the consumers to fully experience the products or services with limited senses, lack of tangibility and sense of presence. Therefore a vital research question can be identified; how multimodal interactions can be used in e-commerce with context awareness, to improve the consumer experience. To address that question, this research aimed to study multimodal interactions, contextual factors and their effects on consumers. A set of multimodal interactions including 3D visualization and hand gestures and related contextual factors such as user, access device were identified in this research. They have been used to develop a multimodal interactions enabled prototype e-commerce framework. Several experiments and user studies have been conducted using the developed e-commerce framework and interesting effects on consumers have been discovered including positive user experience, improved value perceptions, and positive product opinions. Most importantly it has been shown that consumers perceive about 50% increased product value, and they are more likely to purchase when interacted multimodally. Usability Evaluations on the framework showed that users are mostly successful and comfortable in using multimodal interactions. Some technical, social and cultural barriers and challenges for enabling multimodal interactions were also revealed in those evaluations. From the findings of this research, it is suggested that further research focus should be on overcoming the identified technical, social and cultural barriers and bringing multimodal interactions to mass usage in electronic commerce platforms. Also the multimodal interactive e-commerce framework developed in this research can be used as platform to further study consumer dynamics by changing various variables.
- item: Thesis-Full-textSinhala-Tamil statistical machine translation (SMT) for official documents(2018) Farhath, FF; Ranathunga, S; Jayasena, SSinhala and Tamil are declared to be the offi cial lang uages of Sri Lan ka. This requires each government related dissemination/communication to be done in both the languages. Even though the requirement for translation is higher, the number of available human translators is limited. One feasible option to boost the productivity would be assisting the human translators with machine translation output. Here the machine translation output is given to translators to work on by post editing, rather than translating from the scratch. However, Sinhala - Tamil pair does not have any well-performing machine translation system. Therefore, the focus of this research is to develop a machine translation system for short official government documents. This thesis presents two main contributions towards building ‘Si-T a’, the first domainadapted machine trans lation system for Sin hala - Tam il. The first contribution is building the baseline translation system. The second is implementing data pre-processing techniques to improve the translation quality of the base line sys tem. The base line system was built using Moses, a phrase -based stat istical trans lation system. This was the feasible option with the available resources. To improve the quality of the translation, three main approaches were explored. They are: (a) domain adaptation, (b) integration of terminology, dictionary, and name lists, and (c) addressing out-of-vocabulary (OOV) problem using word-embedding-based paraphrasing. In or der to adapt the sys tem for the dom ain of official government documents, different language model design techniques and a data filtration technique were experimented. Under terminology integration, experiments were carried out to evaluate the effect of incorporating bilingual terminology lists to the system. Moreover, a novel data augmentation technique was experimented to generate parallel data using bilingual lists and available parallel data. Further, open domain dictionary entries, as well as a list of person names and addresses were integrated and evaluated. In addition, word-embeddingbased paraphrasing was used along with a novel heuristic-based filtering to address the out-of-vocabulary issue. All the above-mentioned approaches gave an improvement over the baseline, apart from data filtering technique. Yet, all these scores were above the scores of already available machine translation systems for this language pair. Though our techniques/approaches were evaluated only on Sinhala - Tamil pair, they are feasible to be applied to other low-resourced, highly inflectional language pairs.
- item: Thesis-AbstractA Model based approach for cluster traditional rice varieties of Sri LankaSilva, MDRL; Wickramarachchi, NAs a result of the enormous volume of data produced by highly developed modern techniques, focus on clustering biological data has shown a great interest among biologist to detect the underlying patterns in data since the biological experiment itself has failed to identify the hidden information and divergence patterns exist in data correctly. This study aims to (1) assist clustering biologically similar sequences to detect divergence patterns exist in rice genomic data, by developing a program using the model based clustering algorithm based on Chinese restaurant process which was originally proposed to cluster gene expression data (2) focus on nding the performance of calculating the pairwise distance matrix of rice genome sequences based on the 12-dimensional natural vector of the DNA sequence, as the similarity measure in cluster analysis. The developed program based on the proposed model based clustering method was executed on ALFP pro le data set consisting features of 53 Sri Lankan traditional and wild rice varieties in order to identify the genetic divergence among them. Both a statistical and a biological cluster evaluation were carried out to validate the results obtained. Statistical evaluation was done based on the Bayes ratio to measure the tightness of the clusters formed. Biological evaluation was conducted with the help of the domain experts and research work done by the institute of rice of Sri Lanka. The results showed that the proposed algorithm is capable of identifying highly similar varieties of rice showing their divergence patterns. Finding the performance of how well the natural vector method captures the information encoded in rice genome sequences, 10 rice disease resistance genes which belong to three di erent protein families from Rice genome annotation project database were used. The results showed that the pairwise distance matrix calculated based on 12-dimensional natural vector method gives e cient results compared to traditional proximity matrices. It also revealed that the xed length size sequences (sub sequences) which are not greater than the minimum total length of the selected sequences are also highly capable of capturing the encoded information in total length, regardless of the sub sequence length.
- item: Thesis-AbstractRequirement-based policies for electronic message distributionFernando, SD; Wijesoma, WSMany parts of the world endeavor to extend access to more individuals and institutions by means of the electronic messaging system in the ever-evolving Internet. While that messaging system is continuing to incorporate more and more features such as multimedia information, technical barriers in many parts of the world or uncontrolled behaviours of originators hinder such expansion. It is the lack of user-requirement based policies in the messaging system that increases the constraints to such environments. At the same time the need to integrate and coordinate other messaging systems such as fax, paging, postal mail etc. with the Internet messaging system still exists. On the other hand whether constrained or not, users wish to administer their retrieval of messages in particular manners such as blocking unwanted messages, using alternate message delivery media, using strategic message delivery methods especially for lengthy messages, etc., which in turn leads to a formation of requirement-based policies of the messaging system. Currently there are standards governmg message handling, in particular MIME (Multipurpose Internet Mail Extensions), for exchanging mail messages encompassing a multitude of media, such as graphics images, voice, data and full motion video apart from plain text. However, the MIME standard presupposes certain minimum technical capabilities amongst interconnected and participating mail servers and nodes for distribution of such multimedia mail. In particular, the interconnecting channels between mail servers should be of sufficient bandwidth to conduct the large amount of data in MIME messages at 'reasonable' rates and also the nodes must have adequate storage capacity for the same. This requirement (or bandwidth of channels and storage of mail servers for MIME capability prevents users connected to 'under-privileged' mail nodes from enjoying the benefits brought about by multimedia information and messaging. This may be in spite of the end users owning or having access to resource rich machines. On the other hand, with the introduction of MIME capabilities, users are faced with the essential requirement of administration at the server side owing to the fact that the availability of many media types would impose increased abuses, intended or accidental, especially in the face of new inventions of media types. As a strategic solution to this phenomenon, the concept of an adaptive, service-oriented mail server employing a smart approach to routing of multimedia messages in an inter network of mail servers, disparate in storage capacity, performance, network bandwidth and administration, is presented. This concept, described as Hierarchical Actions Transfer (HAT) concept, involves offsetting the bandwidth and capacity limitations or "desires" of a node at a specific level by requesting a mail node a step higher up to oblige to perform services on its behalf (which the former node is incapable of performing or not willing to, due to capacity or bandwidth limitations or self imposed restrictions ). The service requests that are in line with the Requirement-Based Policies are known to the server through a specific configuration mechanism. When delivering messages, the server will first look at the configuration of each connected node and then the message delivery will take place accordingly. It would be inevitably a service-oriented strategy to relieve nodes in constrained environments. An added advantage of the approach is that it implements administrative policies for controlling traffic and congestion arising from MIME mail to a mail server site and the level of MIME services made available to users connected to that site. Further, the technique also integrates conventional and primitive messaging mechanisms, such as postal mail, courier by diskettes/tapes, facsimile and paging, into the realm of electronic messaging. It could be further enhanced by the incorporation of Artificial Intelligence if it could identify patterns of frequent node requests. However, current implementation that uses "send-mail" as the Message Transfer Agent (MTA) focuses only on the manual and semiautomatic configuration of such services.