Browsing by Author "Rajapakse, C."

Now showing 1 - 11 of 11

Android Mobile Malware Detection using Deep Ensemble Machine Learning
(International Postgraduate Research Conference 2019, Faculty of Graduate Studies, University of Kelaniya, Sri Lanka, 2019) Liyanapathirana, C.; Rajapakse, C.
The Android operating system is one of the most used operating systems in the world and has become a target to malware authors. Traditional malware detection methods such as signatures find it impossible to deal with detecting complex and intelligent malware which are capable of obfuscating and repackaging to avoid being detected. There is therefore an increase in the need to have more efficient and intelligent forms of malware detection. Recently deep machine learning and ensemble machine learning algorithms were used to malware detection and classification. Most of the shallow learning models such as SVM, Random Forest etc. had given less accurate results. Hence this research is focused on using deep learning and ensemble methods for better accurate results. Due to its accuracy and intelligence it has become an ideal solution to bridge the gap between traditional classifiers and the intelligent malware. Methodology used was based on opcode, syscalls and API calls in integrated using static and dynamic analysis. Currently, research is mainly being conducted using deep learning techniques to target all or a given malware family. Research addresses several issues related to android malware detection. One such is to proper identification of obfuscated and repackaged android malware packages using the implemented platform. Next research managed to solve one of the major problems faced in dynamic analysis. This is namely the issue of malware going to a silent mode once tested in the sandbox. This problem was also addressed within the research. This paper proposes a methodology which brings an ensemble solution between the shallow machine learning algorithm and deep learning algorithm to create a solution that provides a higher accuracy and performance friendly application to detect and classify malware
An Application of Transfer Learning Techniques in Identifying Herbal Plants in Sri Lanka
(IEEE International Research Conference on Smart computing & Systems Engineering (SCSE) 2019, Department of Industrial Management, Faculty of Science, University of Kelaniya, Sri Lanka, 2019) Azeez, Y.R.; Rajapakse, C.
Sri Lanka has a considerable collection of plant species that have been utilized for generations as medicinal treatments. Knowledge regarding herbal plants is restricted mainly among practitioners in traditional medicine. Available systems studied; had no proper methodology to search information regarding herbal plants, which can be identified through analyzing an image of an herbal plant given. Systematic literature review was done based on herbal plants in Sri Lanka, transfer learning and plant image recognition and two open ended interviews were conducted with traditional medicine practitioners. As main objective of the study, reorganization of Information was done building a technique to enhance capability of identifying herbal plants based on deep convolutional neural networks and image processing techniques which would ultimately assist more locals with identification. Five herbal plant types were chosen to analyze further in detail and the images of the plants were acquired from web and also images photographed via 13MP camera creating a data set validated through traditional medical practitioners. Images were preprocessed and retrained on Inception-v3, Resnet, MobileNet and Inception Resenet V2 based on transfer learning. Algorithm was finetuned using image processing techniques for preprocessing and prototype was tested 5 times reaching highest average accuracy of 95.5% on Resnet for the identification of 5 different plant types. Conclusively, this study enhanced the capability of searching herbal plants by reorganizing the information
An approach to personalize learning using big data analytics for higher education
(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Jabir, A.; Rajapakse, C.
The concept of BYOD (Bring Your Own Device) has gained popularity in studentcentered learning and higher education institutions make significant investments on improving the wireless network to enhance this. Virtual Learning Environment and Learning Management Systems were introduced and personalization of learning becomes the next milestone. The huge streams of data produced by these Wi-Fi networks makes ground for Big Data analytics to identify opportunities in educational environments to adopt personalized learning. The term ‘Personalization’ refers to the tailoring of content and recommending items by inferring what interests a user based on previous or current interactions with that user, and possibly other users. This research proposes an approach to personalize learning on an online learning platform by providing personalized recommendations of educational web resources, comparative feedback and allocate personalized bandwidths based on the concept of deprioritization (lowering priority ranks of heavy users). Concepts of Big Data analytics and data mining techniques will be used to satisfy the objectives. The approach consists of offline phase (modelling phase) and online phase (recommendation /deprioritization) phase. In the offline phase, models will be developed for recommendation and deprioritization separately. For recommendation a hybrid filtering method will be used. k-Nearest Neighbour, a user-based collaborative filtering technique, will be used with correlation based similarity measure with demographic filtering based on demographic classifiers (faculty, year, General/Special/Honors, GPA) to eliminate the cold start problem. To increase the efficiency and accuracy, k-means clustering will be used as an intermediate step to determine usage clusters to group users exhibiting similar browsing patterns and page clusters to discover pages with similar access patterns. For this the access logs of the University of Kelaniya’s Wi-Fi network will be utilized. The parameters for usage clustering would be the timestamp, web resource and category (education, social networking, gaming etc.) whereas the parameters for page clustering would be category and temporal concepts. In the online phase, first the cluster that the current active user belongs to will be identified and k-NN will be applied on that particular cluster to recommend web resources. These techniques also provide the basis for comparative feedback compared to top scorers of the same area of major. For personalized allocation of bandwidth a separate k-means clustering will be performed to identify heavy users during the offline phase. During the online phase deprioritization will be applied accordingly if the current user belongs to the heavy users cluster and there is a heavy traffic in the network. Cross validation will be used to evaluate the models.
A Blockchain-based decentralized system to ensure the transparency of organic food supply chain
(IEEE International Research Conference on Smart computing & Systems Engineering (SCSE) 2019, Department of Industrial Management, Faculty of Science, University of Kelaniya, Sri Lanka, 2019) Basnayake, B. M. A .L.; Rajapakse, C.
Low quality agricultural products are added to the market daily. Over usage of chemicals in the production process, use of uncertified chemicals and mechanisms for preservation and ripening processes, are the major issues that impact on agricultural product’s quality as well as overall health of the consumers. Mechanisms to identify the quality of the agricultural products are highly demanded due to the lack of transparency in the current process. Blockchain technology is emerging as a decentralized and secure infrastructure which can replace involvement of a third party to verify the transactions within the system. The purpose of the research was to implement a Blockchain based solution to verify the food quality and the origin of the agricultural supply chain. A public Blockchain concept was selected instead of a private Blockchain in this study to ensure transparency by allowing any person to access the network. Instances of the smart contract were created for each physical product and deployed to Blockchain network. A Quick Response code which contained the address of the instance, was a reference to the virtual product. All the actors who are involved in the supply chain must be able to interact with the system to achieve the transparency. Each transaction and events related to a product is validated by peers of the Blockchain system. Product ownership was changed for each relevant transaction. A token-based mechanism was used to indicate the farmers’ reputation with their products. Farmers could place a certification request regarding their products and, they can gain reputation tokens for each certification done by peers. A unique Quick Response code was used to identify each product within the supply chain. The proposed system has been implemented as a prototype and validated within the study
Classification of vehicles by video analytics for unorganized traffic environments
(International Research Conference on Smart Computing and Systems Engineering - SCSE 2018, 2018) Arachchi, I.M.R.; Jayalal, S.; Rajapakse, C.
Traffic monitoring is essential for infrastructure planning and transportation. The objective of traffic monitoring is to have an effective traffic management system. Traffic management systems would be effective in well-organized traffic environments, where it has very disciplinary behaviors and less in inefficiencies. But in unorganized urban environments like Sri Lanka, road traffic behaviours are varying from standard structured ways which lead to discompose the traffic management. An effective monitoring system requires short processing time, low processing cost and high reliability. The paper proposes a novel vehicle detection and classification algorithm based on background filtering and re-engineered with suitable changes in order to be applicable to challenging unorganized traffic environments. The solution is successfully classifying vehicles individually and their trajectories in unorganized traffic environments in order to monitor the behaviors of the drivers. The system gives 74.4% average accuracy in vehicle detection and 55% accuracy in vehicle classification while counting each vehicle passed by. We used OpenCV functions for implementing and testing algorithms. Data was collected through pre-recorded video clips from footbridge crossing at Colombo Fort in western province Sri Lanka, for the testing. The ultimate objective of this research was to come up with a best-suited algorithm for vehicle detection and classification (hybrid solution) in unorganized traffic environments which would help to analyze the behaviors of road users. The solution will lead to help reduce unorganized traffic congestions by enhancing the efficiency and effectiveness of traffic monitoring and analyzing systems those are used for intelligent traffic management systems and traffic simulation models.
Evaluating optimal lockdown and testing strategies for COVID-19 using multi agent social simulation
(Faculty of Science, University of Kelaniya, Sri Lanka, 2020) Dunuwila, P.M.; Rajapakse, C.
COVID-19 pandemic has become a major concern due to its rapid spread throughout the world.We can observe some countries are successful in formulating strategies effectively for managingthe transmission of the pandemic, while some countries like USA, India and Mexico are struggling to identify effective policies. Recently, we can observe an increasing trend for COVID-19, surging in the Asian region. The study is based on the question of formulating effective policies for curbing the surge in COVID-19 pandemic by reducing community transmission. While many countries are suffering from the pandemic, it is a critical issue that the policymakers should be concerned with formulating effective policies to address the problem. Computational methods are used to foresee the future by creating a simulation model based on multi-agent methodology since statistical methods require the collection of large amounts of accurate data to train the model which is a challenge, currently. Multi-agent simulation helps in studying macro-level emerging patterns in a complex adaptive system such as a society, by simulating the micro-level interactions of individual entities in the system. A survey and literature review are carried out to collect data on people behaviour, responses for different policies, and social composition. When the model runs, simulated agents such as children, parents, and grandparents will engage in their daily tasks. They will have states of susceptible, infected, or recovered. Based on the testing rate and lockdown day parameters, it identifies different zones as contaminated, buffer, and sterile based on whether any infected people live in that area. The implementation of the model follows an iterative process for improving the validity of the model by comparing simulation results with real-world observations. The validated model can be used for exploring and analysing possible emerging patterns related to community transmission of COVID-19 in the society based on different lockdown and testing strategies such as closing schools and universities, reducing visits to supermarkets by the community, use of public transportation and using aggressive testing and lockdown strategies. The results show that when there are no policy measures taken, the pandemic spreads quickly in the community. When the schools and universities are closed, there is a delay in the pandemic, but eventually, most of the community will get infected. When there are policy measures taken to restrict visits to public places, closing schools / universities and a high percentage of people using private transport, show a slight improvement in controlling the pandemic. However, when aggressive testing and lockdown policies are implemented and carried out, the authorities will be able to control the pandemic within a reasonable period compared to other policies. Further, the implications of the study could be used as a decision support tool for analysing lockdown and testing strategies for controlling community transmission of COVID-19 pandemic.
Instagram sentiment analysis: Discovering tourists’ perception about Sri Lanka as a tourist destination
(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Ranaweera, E.H.; Rajapakse, C.
Today the web has changed from static containers of information to dynamic platforms where users can share digital contents such as blog posts, pictures and opinions in a very simple manner. Especially, the social media is largely getting popular due to the fact that most people prefer to share their feelings, thoughts and memories of their daily activities in social networks. One of the most common types of posts in social networks is opinion related posts. Moreover, social network users tend to seek opinions of others before purchasing a product or getting a service. Social media plays a revolutionary role in travel and tourism industry. With the increasing use of social media, tourists not only consume tourism products and services but also prefer to share their experiences with others in the forms of textbased opinions, comments to other’s posts, pictures with descriptions, ratings, etc. Current statistics available with Sri Lanka’s tourism authorities do not reveal whether tourists are happy with the services received during their visit and they have no information regarding common issues that the tourists have to deal with when they are in Sri Lanka. However, reading and analyzing all these online posts is not practically feasible due to the enormous time and human resource that would be required. The objective of this research is to identify how social media contents could be used to extract valuable and meaningful information to develop and promote travel and tourism industry in Sri Lanka. Our approach is to adopt Sentiment analysis techniques to analyze the text-based contents shared by tourists on Instagram, which is a popular social networking site among tourists worldwide, to determine the overall perception of tourists about Sri Lanka as a travel destination. Photo descriptions and user comments are collected, using special keywords related to tourism in Sri Lanka using an online tool and, in the first phase of the research, sentiment classifier with support vector machine algorithm will be develop to identify sentiment polarity of posts. Furthermore in the second phase feature analysis model will be developed through which positive posts with feature words will be used to identify tourists who recommend Sri Lanka to others or potential tourists who plan to visit/revisit Sri Lanka. Moreover, feature categorization method will be used to identify the key areas that require improvements to offer a better service to tourists through negative sentiments.
Investigation on the Adaptation of Business Intelligence and Analytics in Sri Lankan Supermarket Sector Organizations
(Faculty of Computing and Technology, University of Kelaniya, Sri Lanka, 2016) Ranasinghe, P.; Rajapakse, C.
In the era of the fast moving technology, every company and firm try to get the competitive advantage over the other firms in participating actively in the competition of the market. As the businesses get larger and distributed, the data collected each day rise from megabytes to terabytes each day. Increasing amounts of data give the opportunity to companies to use analytics to understand the hidden patterns in the data they collect and harness valuable business insights to gain competitive advantage. We present the results of a study on the adoption of business analytics in the Sri Lankan supermarket industry conducted to evaluate their readiness to use the state of the art business intelligence technologies available today. The main reasons for selecting supermarket sector over another industry is due to the nature of transactions, volume of transactions, high competitiveness, requirement for analysis, proven ability to get good results by using business intelligence techniques, loyalty card systems, the huge volume of customer data and transaction data collected each day are significant. Furthermore, the supermarket sector exhibits a huge competition among the three leading conglomerates and one appearing conglomerate. In order to gain competitive advantage these companies should understand the patterns hidden among these data such as the behaviors of their consumers and the trends in the market. However, it is unknown whether they are adequately and successfully adopting the business intelligence technologies for competitive advantage even though they have sufficient data assets. Through a through literature review the factors were identified that should be considered related to assessing the readiness of a company for business intelligence and analytics. A questionnaire was made based on the review of literature elaborating the seven factor model referred as “BI readiness Assessment” which can be used to determine the states of various issues related to organization’s ability to utilize BI. “BI readiness Assessment” describes seven readiness factors, which they have also referred as potential barriers to the ability to deploy BI. We used this model, originally developed for the context of the United Kingdom, for the context of Sri Lanka and thereby evaluated the Sri Lankan supermarket sector organizations in terms of the seven factors of BI readiness to understand the overall readiness and adoption of analytics. The research was designed as a case based qualitative research in which all identified leading supermarket sector conglomerates were analyzed as cases. The four leading conglomerates were identified based on their variety as well as annual turnover and the government owned Supermarket Company. Extensive open ended interviews with IT managers and the heads of BI departments of the respective organizations based on the questionnaire developed were used to collect data to develop the cases. The questions covered the seven factors of the model, in order to verify each a set of open ended questions were defines and gathered details through answers. Qualitative data that had been received is translated into useful information through context analysis. The questions under each seven factors considered are translated to numerical values and given the ranks according to Mann–Whitney U-test. The ranks for the main seven factors are derived then with the weighted average for each point. By comparison of the ranks the results are derived. In Sri Lanka there are only five main supermarket chains and all of them are interviewed, and gathered details. So the coverage is 100%. Four of them belong to private sector two of them are part of the big conglomerates, the other one is owned by the government of Sri Lanka. In the study we came across with different levels of usage of business analytics; Firm 1 is using query processing for analytics, Firm 2 is using a business intelligence tool, Firm 3 is having a stable ERP culture in which they perform analytics as well as an ongoing project to implement a BI system based on Hanna, Firm 4 is having a strong infrastructure design but still in the process of implementing the infrastructure and the Firm 5 shows as an outlier, which doesn’t uses any business intelligence or analytics, the firm is still in the process of getting point of sales systems all outlets, from all details of stores all over the country nothing is collected and put into a common system or linked. The summary of our analysis based on the seven-factor model is given in Table 1. As per the methodology I used the qualitative analysis As shown by the summary of the analysis, it is clear that all large-scale supermarket sector companies use and utilize BI and analytics for a considerable scale, hence their sufficiently ready for the analytics world. However, the world is now moving in to the world of “big data”, which is largely characterized by unstructured data and, investigating the readiness of these companies for analytics in the big data world would be an interesting future research that extend this study. Our study indicates that except “Firm3”, the other firms are not ready to analyze big data yet.
Real-time big data video analytics for unorganized traffic environments
(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Ranaweera Arachchi, I.; Jayalal, S.; Rajapakse, C.
Traffic on local roads has reached such a level that it is necessary to address the issue of traffic congestion and seek complex transport solutions for the city. Increase of the number of vehicles on the road becomes one of the key reasons for increasing traffic congestion. Traffic congestion is associated with massive financial and manhour loss and therefore attempts to alleviate this has been of keen interest. The basis of almost all those approaches is traffic monitoring and analysis, leading to having an effective traffic management system. Most traffic management systems are applied in well-organized traffic environments such as highways, where driver discipline is high. But in unorganized urban environments as seen in Sri Lanka, road traffic behavior vary from the accepted standards. Driver and pedestrian indiscipline cause huge traffic congestions in urban areas. Hence in such a scenario, a system that monitors road traffic on different traffic environments is very useful. There are several existing techniques such as Magnetic Loops, Microwave RADAR, Infrared Detectors, Ultrasonic Detectors and Camera Based Systems. Traffic monitoring systems require short processing time, low processing cost and high reliability. Therefore, according to the literature, camera-based monitoring is the best-suited technique for traffic monitoring. Real-time video analytics are part of a centralized approach to modern traffic management which is defined as computer vision-based surveillance that provides algorithms for object detection, tracking, classification and trajectory analysis using real-time traffic surveillance video. It usually uses roadside cameras (CCTV) to obtain traffic information and transmit it to central servers, exhibiting real-time operability of big data. In this study, several approaches and algorithms for moving object detection, based on temporal differencing method, optical flow method, background filtering are compared and a novel real-time vehicle detection and classification algorithm based on background filtering will be proposed and re-engineered in order to be applicable to challenging unorganized traffic environments. The solution will classify vehicles individually and their trajectories in real time in unorganized traffic environments in order to analyze the behaviors of the drivers as well as pedestrians on the road. We use OpenCV which is a library of programming functions mainly aimed at real-time computer vision, for implementing and testing algorithms. Data will be collected via pre-recorded video clips from Kiribathgoda junction in the western province, for the testing purpose and real- time CCTV surveillance video is going to be used as the input for implementation. A comprehensive data analysis is required to be conducted to address the higher processing requirement of such videos. The solution will be validated for performance subsequently. The final objective of this research is to come up with an optimum algorithm for vehicle detection and classification in unorganized traffic environments which would help to analyze the behavior of road users. The solution will lead to reduced traffic congestion in the country by enhancing the efficiency and effectiveness of traffic monitoring and analyzing systems.
Urban traffic simulation using agent-based modelling: A study in the Sri Lankan context
(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Amarasinghe, U.G.L.S.; Rajapakse, C.
Traffic congestion is a crucial issue affecting the quality of life of individuals all over the world. In a country like Sri Lanka where the traffic is mostly unorganised and mixed, traffic congestion occurs due to various reasons such as the volume of traffic exceeding the capacity of the road, road accidents, temporary closures of roads due to constructions, as well as the behaviours of pedestrians and drivers. For example, careless lane changing behaviours of drivers and the bad practices of crossing streets of pedestrians account for a larger portion of urban traffic congestion every day. Due to the significant impact of traffic congestion to economic growth, various approaches have been taken by researchers and administrators to reduce the urban traffic congestion. Some popular approaches to solving this issue includes infrastructure development, introducing new traffic rules such as changing peak hour traffic plans in cities, as well as imposing heavy duties on vehicle imports to reduce the growing volume of vehicles on roads. However, despite all these attempts, the traffic congestion remains a serious issue in Sri Lanka. Traffic simulation is one of the most effective tools for the testing of traffic solutions and finds the reasons causing traffic congestion. Traffic conditions are different from region to region due to different factors: traffic laws, vehicle types, drivers’ and pedestrians’ behaviours. Therefore, researches have been done by focusing on modelling traffic simulators considering those factors specific to particular regions. We propose the Agent-Based Modelling and Simulation (ABMS) approach, which is a popular computational research method based on swarm intelligence to study complex social and economic systems, to model a traffic simulator simulating mixed traffic conditions in Sri Lanka which is an unaddressed area of research. In this approach, individual vehicles and pedestrians are modelled as software agents who have a set of individual (i.e. micro level) behavioural rules. When these agents are put together, they behave as the vehicles and pedestrians behave in the real world interacting with each other giving rise to emergent macro-level patterns, which we call traffic congestions. This study aims at modelling vehicle following behavior, seepage behaviour of vehicles and pedestrian’s behaviours at un-signalized crossings. We use the ABMS environment called NetLogo to develop our simulator and Kiribathgoda junction in Western Province, Sri Lanka as the test bed. Data collected from there will be used to calibrate the model with accurate parameter values. Macroscopic statistics such as the rate of traffic flow, average speeds and queue time will be used to validate the model by comparing data from real traffic situations with model outputs. The ultimate objective of this research is to come up with a cost-effective decision support tool for administrators and policy makers to understand various reasons behind congestion in unorganised mixed traffic environments in Sri Lanka, apply and evaluate different traffic control strategies and thereby to make better-informed decisions to control urban traffic congestion in Sri Lanka.
A zero configuration protocol stack for device-to-device communication in a private Wi-Fi network
(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Wickramarathne, I.Y.; Jayawardena, B.; Rajapakse, C.
Traditional wireless communication involves many Internet-based services, which makes the availability of an Internet connection mandatory to complete the communication. Furthermore, when introducing new devices into the network, preconfiguration, authentication and granted access rights are must to have. This research proposes a stack of protocols for discovering peer nodes, establishing connectivity with them and handling communication between those peers in a Wi-Fi (Wireless Fidelity) network without Internet connectivity as well as configuration and authentication requirements. In this new protocol, all the nodes in the network becomes objects and the user selects appropriate communication channel based on the communication requirement. It could be voice calling, screen sharing, chatting, video calling, file sharing, data transmission, etc. Since the protocol uses local area network, network traffic is not going out from the local router to the Internet. The protocol’s security mechanism is based on different instances. The user is allowed to define his/her own security definition to his/her communication. The protocol supports to network security, application based security, and group based security with encryptions. The research is based on the build and test approach where incrementally developed components of the protocol stack are tested on different Wi- Fi security platforms and device platforms and fine-tuned for minimum bandwidth consumption and data losses. Protocol stack is being developed in accordance with several layers of the TCP/IP (Transmission Control Protocol/Internet Protocol) model such as application, transport, network and physical. Simple chat application is built with all the proposing components and algorithms in order to proof of the concept. Our ultimate objective is to apply this new communication protocol to the IoT (Internet of Things) environments. Since protocol supports to any OS (Operating System) platform and enable Wi-Fi communication with any device without any configurations, this can be used as the core communication protocol used by the devices present in an IoT environment.