ITRU - 2015

Permanent URI for this collectionhttp://192.248.9.226/handle/123/14727

Browse

Recent Submissions

Now showing 1 - 11 of 11
  • Thumbnail Image
    item: Conference-Abstract
    Effect of mutation and effective use of mutation in genetic algorithm
    Dassanayake, VPC
    Author tries to analyze the effect of genetic algorithm when mutation rate is selected randomly compared to fixed mutation rates or/and adaptive mutation rates. The results shows that though it is not always possible to get comparable results using randomly selected mutation rates, it is possible to obtain the required range with less number of trials than using fixed mutation rates for this nature applications
  • Thumbnail Image
    item: Conference-Full-text
    States Prediction of Web Services Using Hidden Markov Model
    Prasanga, RKM; Wijesiriwardana, C; Weerasuriya, GT; Fernando, S
    Over the last few decades, service oriented architectures, in particularly web services, have grown in popularity in the context of enterprise level application integration. As a result, most of the enterprise level software systems tended to be developed with a flavor of web service components. However, like all other distributed software technologies, web services also fail. Therefore, proper mechanisms and tools to handle system failures are vital to avoid such exceptional behaviors. To address that problem, this paper investigates a state prediction mechanism for web services using Hidden Markov Model (HMM). This approach is capable of predicting the future exceptional behaviors of the web service by analyzing and identifying the error patterns generated by long-running web services. This research can be further extended with an automated system input to determine the system state.
  • Thumbnail Image
    item: Conference-Full-text
    Sinhala Handwriting Recognition Mechanism Using Zone Based Feature Extraction
    Dharmapala, KAKND; Wijesooriya, WPMV; Chandrasekara, CP; Rathnapriya, UKAU; Ranathunga, L
    Identification of Sinhala characters is considerably more difficult than other wide-spoken languages because of the complex shapes and similarities that are present within characters. With the addition of modifiers to the core characters, the recognition becomes increasingly more difficult. Most of the present systems only address the identification task of core characters which has potentially less real life applicability. The proposed solution tries to identify characters with or without touching and non-touching modifiers which can be effectively used in multiple applications.
  • Thumbnail Image
    item: Conference-Full-text
    Semi-supervised learning framework for knowledge extraction in Cricket domain
    Fernando, CCMT; Cooray, WAVS; Indeewara, TGH; Jayasinghe, DSK; Fernando, S
    Today’s web is overwhelmed by the data and it is continuously growing, therefore processes of information retrieval and analysis have become tedious tasks. Although many machine learning approaches have been applied to mine these data, many of them are hardly succeeded in their approaches because they have restricted themselves into targeted or downloaded databases. Growing web can not simply be classified or mined by using static knowledge base, system has to grow with the web. Therefore, a system that can mine while learning from the mined data, is required. This paper proposes a framework that acts as a learning model to derive information by building relationships between different entities in online content by relying on few seeds being fed to the system at the start. Couple of extractors are used to derive facts based on their mutual correlations. Those facts have been occupied an ontology to generate new relationships and entities as candidates. A query system has been embedded to the miner to enable querying the knowledge base to retrieve appropriate outputs corresponding to a particular query. The system has evaluated against the cricket online sources.
  • Thumbnail Image
    item: Conference-Full-text
    Performance Evaluation of the User Interfaces in Software Programs using Eye & Gaze Tracking
    Athukorala, D; Ganhewa, H; Jayasinghe, U; Karunarathne, D; Ranathunga, L
    There are many software programs developed by many software developers in these days. The people who are going to interact with them need some time to learn about how to do their tasks using those software programs. Some of these software programs need more than expected time to learn about using them because of poor design of the User Interfaces. User Interfaces are the main windows which are used by the people to interact with the software programs, because of that they must be designed in better way which attracts the attention of the user. Therefore in Human Computer Interaction principles, there some principles for User Interfaces design effective User Interfaces to attract attention of the user to the program. When designing software programs it is necessary to find hotspots on the User Interfaces where the user gives more attention than others. This paper proposes a method of performance evaluation of the User Interfaces of the software programs by testing the software program with actual users of the software program and tracking gaze point of users with the time. By using that information the software developers can find the hotspots of the User Interfaces and include the most important things to the user in those areas. The software developers will be able to develop the User Interfaces of their programs attractive, efficient & effective manner which get more attention of the users of that software program.
  • Thumbnail Image
    item: Conference-Full-text
    HTML5 Based Email Client with Touch Enabled Advanced User Interface for Tabs and Tablets
    Madhuka, GPD; Dias, GV
    The Internet has become as a very powerful platform that has changed the way we do business and the way we communicate. E-Mail is an essential tool for both cooperative world and individuals for communicating. Web-based e-mail systems have become more popular among the internet users with time. Internet and the devices that we use to access the internet are rapidly changing time to time. Among the recent evolutions the most significant of them is HTML5 incorporated mobile technology with hi-tech devices like tabs, iPads and tablets with touch sensitivity. However, the major problem occurs when different levels of resolutions arisewithmodern devices. Some devices support touch, multi-touch, gestures, keyboards and stylus. User-interface of email web clients hasnt improved in the way hi-tech devices evolved. Handling different types of interactions depends on the device and way that user handles it. We have developed HTML5 off-line supporting web-based UI for e-mail system to overcome this issue and to provide a highly user based interactive, responsive and efficient process even in slow network connections. Our approach is based on HTML5 features and client side on java-scripting. Our system is capable of running on a browser without installing any plug-ins. Depending on the device resolution and user interaction (one finger touch/ both hands or external keyboard) email client has provisions to transform the web UI to give better interaction for the user and email system.
  • Thumbnail Image
    item: Conference-Full-text
    Effect of Mutation and Effective Use of Mutation in Genetic Algorithm
    Dassanayake, P
    Author tries to analyze the effect of genetic algorithm when mutation rate is selected randomly compared to fixed mutation rates or/and adaptive mutation rates. The results shows that though it is not always possible to get comparable results using randomly selected mutation rates, it is possible to obtain the required range with less number of trials than using fixed mutation rates for this nature applications.
  • Thumbnail Image
    item: Conference-Full-text
    Automatic Glaucoma Detection by Using Funduscopic Images
    Atheesan, S; Shanmugarajah, Y; Ajanthan, T; Ranathunga, L
    This paper describes an automatic system to identify glaucoma disease from funduscopic images by using digital image processing. Glaucoma caused by increase of pressure in eye and damages in optic nerve. Glaucoma tends to be grown and may not show until final stage. Through this system, doctors can easily identify patient’s condition quickly and do treatment. Rural people also will get advantage through this system. Glaucoma is identified through cup to disc ratio (CDR) calculation and orientation of the blood vessels in this system. For that Optical disk’s inner circle (cup) and outer circle (disc) is extracted. From that radius is calculated. The outer and inner circles are extracted by using average and maximum grey level pixels respectively with the use of histogram. Then find contours and draw circle which is best fitting the contours. The radius of cup and disc are found. After calculating CDR, the abnormal image can be found if CDR exceeds a particular threshold value. Otherwise it is normal image. The system extracts the blood vessels and through the orientation of the blood vessel glaucoma is identified.
  • No Thumbnail Available
    item: Conference-Full-text
    Analyzing the Healthiness of an IT Project Using Self Organizing Maps
    Thilakarathne, HHTD; Wellage, CH; Rupasinghe, JAPNS; Nimantha, KC; Karunaratne, PM; Fernando, S
    This research paper discusses a solution for a problem we have identified that IT companies face when managing their projects. Project managers often find themselves in a tough situation when deciding the current status of a project and making decisions based on the evaluation. But similar situations have happened earlier in other projects and the knowledge about the measures that were taken at those situations and their effect on the success of the project can be used to evaluate similar situations in new projects. Our approach in this regard is analyzing the past data of IT projects using different machine learning techniques to identify the major factors that have affected the success of a project, understand how strongly each factor is bound to success and then training a model with the data. Where it can be used to analyze situations that arise in new projects and identify how like the current situation is to lead the project in to a success or a failure. The machine learning technique we have likely used in this study is Self- Organizing Maps (SOM) and the system was implemented using Python language
  • Thumbnail Image
    item: Conference-Full-text
    A Mobile Application with Augmented Reality to Enhance Sinhala Learning Experience for Children
    Wickramasinghe, WAUYS; Gunasekara, HDA
    Augmented Reality has gone from science fiction dream to possible emerging technology in the modern world. This futuristic technology has the ability to blend the real world with digital world. This paper presents an Augmented Reality based mobile solution for children to enhance their Sinhala alphabet learning experience. With this tool, children can interactively learn letters, pronunciation and usage of letters in real world scenarios. Generally, teaching preschool children could be difficult since they only focus on an activity for a short period of time. Therefore this solution mainly focuses on adding fun and entertainment while addressing cognitive skills of the children. This application adds totally new learning experience for its users. A survey done with pre-school children and teachers proved that using Augmented Reality based solution with latest technologies are very helpful for learning things and teaching things efficiently. Most students requested to use the Augmented Reality application repeatedly. Teachers showed high level of enthusiasm and over 85% of the respondents gave extremely positive feedback
  • Thumbnail Image
    item: Conference-Abstract
    Task allocation in component based global software development
    Wickramaarachchi, D; Lai, R
    Global Software Development (GSD) has recently evolved and embraced by the competitive software industry today. The major attraction for GSD has occurred due to greater availability of human resources in distributed zones at low cost and advancement of communication technology. However, realization of expected benefits is not good as predicted since it involves some hidden cost. Therefore, obtaining the maximum benefit from GSD is a new challenge in this discipline. Since communication and coordination cost is a dominant factor in hidden cost, minimization of it will help to achieve the objective. Proper task allocation is particularly important since it can reduce communication and coordination overheads. There is no proper task allocation method for Component Based GSD. In this paper, we present a module formulation mechanism based on Component Based System Specification (CBSS) written in Unified Modeling Language (UML) identifying component dependency through interfaces. Further, it suggests allocation of work to different sites considering site dependency characteristics.