Androsafe: Online malware analysis with static and dynamic methods

Abstract – With an estimated market share of 70% to 80%, Android as becoming the most popular operating  system for smartphone and tablet. Cyber criminals naturally expanded their various activities towards Google’s mobile platform.An additional incentive for mobile malware authors to target Android instead of another mobile platform is Android open design that allows users to install the application from a variety of sources. “Androsafe” is an online malware analysis tool which can analyze malware in an
isolated environment without any damaging to the mobile device by using both existing and new anomaly based and behavioral analysis. Through this combination, we can analyze a large number of malware families because some malware families may only perform signature base or behavioral. Then the sandboxes based on signature will not have analysis malware families that only perform a behavior and the sandboxes based on behavior will not analysis signaturebased malware families.“Androsafe” sandbox will be hosted in the Honeynet Project’s cloud. Dynamic Analysis will be queued and run in the background, and an email which contains malware analyzing report will be sent to the user when the analysis is over. This method is very efficient more than offline kernel and app base sandbox.

Authors-Krishnadeva Kesavan, Chethana Liyanapathirana , S.A.W.S Sampath , Y.M. Sureni Koshila,Chamod Premarathne ,Sahan Wanigarathna, Chamira Priyamanthi Nawarathna, Prabhath Lakmal Rupasinghe

View Full Paper

Database Optimization Model to Reduce Application Delay Relates to Mobile Devices

Abstract  – In the era of information technology, the  smartphone has become an integral part of human lives. A variety of smartphone applications is available to provide different
types of services required by the users. In order to execute installed applications, the mobile devices consume more memory and processing power. Hence to collect, store,
transfer and combine massive amounts of resources. As lifestyles are becoming fastpaced, a slow and inefficient database applications are be deemed unacceptable.
However, exploiting the information contained in these archives intelligently turns out to be relatively complicated, and it becomes harder to work with these data when it starts to evolve. It will be almost impossible to handle or access a massive amount of data, so it requires optimization method to reduce access time. This paper will discuss vital matrices such as access time, insert time, and delete time and execution time while increasing the performance of the database. Authors will focus on utilizing an Optimize indexes using Profiling to capture cost required to manipulate data. The outcome of this research work will improve the efficiency of the applications by minimizing the response
time. The test results emphasize a comprehensive evaluation based analysis related to the proposed model.
Keywords – database optimization, query optimization, indexing, profiling.

Authors-J.A.D.C. Anuradha Jayakody, D.S.A.Kandawala, K.D.D Chathurangi, N.T Mapa, T.S Dissanayake Department of Information Systems Engineering, Sri Lanka Institute of Information Technology (SLIIT),

View Full Paper

A Sentiment Analysis and Classification Algorithm  Utilizing an Independent Term Matching Scheme Sensitive  to Negations and Word Count Patterns

Abstract  –The “Sentiment” or “Emotion” contained in a short comment particularly when expressed by a  customer in relation to a product or level of service provides valuable feedback to an organization on customer satisfaction enabling timely actions to resolve current and emerging issues of varying degrees of importance to the business. The ability to measure the “Sentiment” and particularly its “degree” allows the severity of the feedback to be quantified. Independent Term Matching Algorithms (ITMA) while allowing for the
calculation of an overall sentiment score for a comment made up of several emotion bearing words or phrases typically do not take into account negations in the comment and therefore miss important meaning contained therein [3]. Typically these algorithms cannot also be trained to recognize patterns within the comments and so are unable to classify comments as to their inherent content. In light of this, we propose a novel algorithm  that allows rating multiple emotion is bearing words in a comment using the ITMA while also accounting for negations. This is effected by modifying the sentiment score of the emotion bearing word occurring subsequent to the negation so that its value is adjusted in a direction opposite in polarity to the affected word. The proposed model is also able to classify comments according to their content so they can be directed to the relevant authority for necessary action. This classification is vital as it allows an assessment of the severity of the issue via its sentiment score while also allowing steps for its resolution in a timely and effective manner by individuals with the relevant domain expertise. We use a process of Association Rule Discovery  to find groups of words that are likely to occur together in a comment to facilitate the classification. This allows the generation of word combinations or word sets that are used in training a Naive Bayes Classifier. The rules are selected based on support, confidence and coverage criteria so that only those that are meaningful to the classification are chosen. This process of keyword generation and association rule prioritization allows us to classify a diverse set of comments relating to a number of sources and a range of issues emanating from a varied
group of customers concerning sentiment, its  degree, and its content.

Authors-Dr. Asoka Korale, C.Eng. MIET, Chanuka Perera, Dip. ,ABE(UK) Eranda Adikari, BSc., C.Eng. MIESL,Nadeesha Ekanayake, B.Sc

View Full Paper

A Light Weight Provenance Aware Trust Negotiation Algorithm for Smart Objects in IoT

Abstract  –Internet of Things can be considered as the next  big tide which advances towards the ICT realm. Many research communities have shown enthusiastic interest towards the variety of research topics which has been emerged into a discussion related to this novel concept. The research taxonomy of IoT is built upon several  key pillars by considering its Complexity, Heterogeneity, and Versatility nature. Among these, security related research challenges can be considered as a key impacting domain. This
particular research has been conducted with the special consideration towards Trust Negotiation among smart objects in order to satisfy provenance related criteria. Therefore this paper has suggested a light –weight, less complex , comprehensive encryption algorithm by applying shuffling techniques in order to satisfy the origin identification.
Keywords— Internet of Things (IoT), Security in IOT, Provenance, Trust Negotiation, Light – Weight Encryption

Authors-J.A.D.C. Anuradha Jayakody, Lakmal Rupasinghe, N.T Mapa, T.S Disanayaka,D.S.A.Kandawala, K.D.Dinusha Chathurangi , Krishnadeva Kesavan
Department of Information Systems Engineering, Sri Lanka Institute of Information Technology (SLIIT), Malabe, Sri Lanka.

View Full Paper

The Simulation – Based Solution to Detection by Collectively Decomposing Factors of IMU and Image Process Data in the Indoor Environment

Abstract -Existing indoor navigation system face with many different technical and usability problems  because of the localization. In this paper presents indoor navigation simulation-based solution that can apply to the real implementation. The proposed system has  several benefits and has the potential to increase the usability of the scheme.
Navigation is the process of monitoring and controlling the movement of and item from an origin to a destination along the path. Navigation system provides reading monitoring
and updating the movement of one’s position and guiding by intelligible visuals. Audible or Tangible means while she is traveling on an intended route. In this simulation movement of the user is shown by the navigation and based on the simulation human movement can identify the path, location, the remaining distance to their final destination. Moreover, this will be critical for the user to determine their exact path. Any number of user can use this at the same time,and it has the facility to simulate the crowd sourced.
Environment. Further, this research work focuses database optimization on reducing access time as an optimized solution.Based on the simulated results authors plan to develop an algorithm that facilitates to indoor navigation with localized information and plan to discuss test results with evaluation.
Keywords- IMU, REST, ath_H_ID, ImgD ,Labelname(LN),PlaceDescription(PD),StartingX(SX),EndingX(EX),StartingY(SY),EndingY(EY),FloorID(FID),BuildingID(BID).

Authors-J.A.D.C.Anuradha Jayakody ,R.G.P Dulshan, D.C Meegahawatta, H.M.K.S.B Herath,L.P.A.D Pathirathne Department of Information Technology, Sri Lanka Institute of Information Technology (SLIIT), Malabe, Sri Lanka

View Full Paper

Predictive Analytics with online data for WSO2 Machine Learner with the support of Ensemble method

Abstract-Different types of malware prevail in a world of innumerable computer users who constantly struggle with threats from malware sources such as the internet, local networks, and portable drivesand so on. Security specialists and malwaredevelopers have been at a tug of war with eachother as security specialists try to use all possible
techniques to overcome the problems arising frommalicious software while malware developers try tobypass these implemented security measures. Itcan be observed from records that each day, nearly1 million new malware threats are released to theworld. Therefore it is evident that there is anindispensable requirement of a proper malware
identification mechanism. Typically, antivirus software is based on a signature definition system which keeps updating from the internet and thus keeping track of known malware. While this was sufficient sometime back, it does not cater to the current requirement of identifying malware. Due to the advancement in technology, malware developers have been able to create malware that is able to obfuscate themselves thus giving birth to polymorphic malware. In this study, we closely observed the behavior of malware, tried to understand how they work, their different types,dissemination of malware and detecting/defending mechanism in order to contribute to the process of security enhancement and came up with the solution of integrating Machine Learning to our current problem.Keywords: Malware family classification, predictive analytics, HTTP Streaming data

Authors-Lakmal Rupasinghe , Heshani Herath , Ishani Pathinayake , Ashani Diaz ,Indujayani Karthigesu Sri Lanka Institute of Information Technology, Sri Lanka

View Full Paper

NExT-Plugin for MetaTrader 5 based on Recurrent Neural Networks

Abstract-This paper reports evidence that a recurrent neural network model applies to the prediction of foreign exchange rates in the Forex Market. In present human brain power has been used to perform the complex process. It essentially needed an effective and efficient method to replace this exciting procedure in order to improve performance and support time serious analysis.Historical and Streaming Time Series data and technical indicators, such as the impact of economic events, are fed to the neural network to
capture the underlying “rules” of the movement in currency exchange rates. The variation of currency pair EUR/USD are forecast by the trained recurrent neural network. The results presented here show that without the use of extensive market data or knowledge, useful prediction can be made, and significant profits can be achieved with reliable predictions. Further research on improving the performance using distributed computing is also conducted. However,the experiments show that with the efficient market it is not easy to make profits using technical indicators or time series input neural networks. After presenting the experimental results, a discussion on future research concludes the paper.
Keywords—forex; recurrent neural network;MetaTrader; time series; forecasting;

Authors-H.A.C.J Hettiarachchi, L.I Hettiarachchi, M.M.N.M.Mullegama and R.K.T.D Ranaweera, L.Rupasinghe, D. Kasthurirathna

View Full Paper

Mıtıgatıon of The Evıl Influence of Internet Use On Teenagers in Srı Lanka

Abstract-  This is the information era. Whole the world is considered as a single village alias the concept of  the global village has been accepted. Because each and every country is connected via Information and Communication Technologies (ICT). Specially, it is the Internet which is a connection of millions and millions of computers or servers
throughout the world. It is said that more than 100 countries are connected into exchanges of data, information, news, knowledge, opinions and many more over the Internet. So, the Internet plays a main roll on the concept of globalization. Hence, the use of Internet is experienced by everyone everywhere and every time. The Internet has divided the society into two portions called “Digital Natives and Digital Immigrants”. With the rapid development of ICT, there is a rapid trend of use of internet among the Digital natives alias the young generation in Sri Lanka, especially among teenagers.Digital Natives’ this new trend of Internet use has brought out another bad trend of evil. So,this evil influence of internet use on teenagers was surveyed.The survey has proven that the evil influence of Internet use has affected the teenagers’ physical and mental health which causes to raise many social and economic issues.After the study of the research data, the main issues as well as the causes for the evil influence of internet use, for the Mitigation of the evil
influence of Internet use on teenagers, suggested recommendations are as follows:To enhance the parents’ observation on the teenagers’ internet related activities.To improve the awareness on issues arise via the Internet use, among the teachers, parents, and the teenagers.To enhance of the parents’ and teachers’guidance to the teenagers on their internet related activities.

Authors- T.G. Premarathne, Yasas Mallawaarachchi

View Full Paper

A Comparative Analysis on Technologies Available for Localization in Indoor Navigation

Abstract-  Indoor positioning systems have become very popular in past decades. These systems have been successfully used in many applications such as asset tracking and inventory management systems.This paper provides an overview of the existing indoor positioning solutions and attempts to classify different techniques and its usage for the localization. Two typical localization methods are addressed in this research work and studied estimation schemes of triangulation and fingerprinting, so analyzed in detail since it is used in the indoor environment. Furthermore, examine the set of performance matrices, and apply evaluation methods to survey some existing systems. Therefore performance comparisons include accuracy, precision, complexity, scalability ,robustness, and cost. The experiment result discusses the above mention performance matrices.
Keywords – Indoor navigation, RSSI, RFID, BLE,Fingerprinting

Authors- J.A.D.C.Anuradha Jayakody , Shashika Lokuliyana ,K.Lakshmi Lashini ,S.A.Rangala  Information Systems Engineering, Faculty of Computing, Sri Lanka Institute of Information Technology, New Kandy Road, Malabe, Sri Lanka

View Full Paper

ClusterMal: Automated Malware Analysis with clustering,anomaly detection and classification of existing and new behavioral analysis

Abstract-ClusterMal is an automated malware analysis system based on the Cuckoo Sandbox technology. Since malware is the Swiss-army knife of cyber criminals and any other adversary to an organization, in these evolving times detecting and removing malware artifacts is not enough: it is vitally important to understand the behavior, context and motivation and
goals of the breach. The Cuckoo Sandbox plays a vital role in analyzing malware but has no clustering feature. As a solution, we propose ClusterMal, which is capable of clustering, anomaly detecting and classifying of existing and new behavioral analysis through machine learning. In the proposed method when a suspicious file is thrown at ClusterMal it
throws it into several different environments (VMs with different OSs installed) and a detailed result is outlined. If no similar behavior is observed, a new cluster is created along with the report and a signature for future usage. In practical implication, the module is integrated into Cuckoo for command line interactions with providing high reliability and
Keywords — Malware Analysis, Clustering, Anomaly Detection, Behavioral Analysis, Sandbox Technology,Machine Learning

Authors-Prabhath Lakmal Rupasinghe , Krishnadeva Kesavan, Sripa Vimukthi Bannakkotuwa, V.V.YY. Wickramanayake , M.P.D.H De Silva , J.M.D. Fernando , K.K.K.K. Sampath Department of Information Systems Engineering, Sri Lanka Institute of Information Technology (SLIIT),Malabe, Sri Lanka

View Full Paper

Reduction of Power Losses in Low Voltage Distribution Network by Improving the Distribution Line Connections

Abstract – This paper presents attractive technical and nontechnical loss reduction methodologies in an existing Low Voltage (LV) distribution feeder by using refurbishment techniques. Further, this study describes power losses due to low voltage distribution systems and emphasizes the effects of nonstandard workmanship. Results were obtained on a selected LV distribution feeder in Jaffna peninsula by using computational tools and reconfigured results were compared with the readings taken before refurbishment.
Keywords – Loss Reduction, Technical losses, Non-Technical Losses, Electrical Distribution.

Authors-Eng. K.Lingaruban, Prof.M.A.R.M.Fernando, Dr. A.Atputharajah, Eng. D.K.P.U.Gunathilake

View Full Paper

Engineer adoption of new technologies to boost Social Development; A Perspective on Digitalization of Terrestrial TV Broadcasting in Sri Lanka

Abstract – Seeking continuous improvement in life style possibly would have been the only objective that  human race never change across the globe, while others do evolved along time and geographical dimensions.

Authors- Kumudu wasantha Muthuthanthrige, B Sc (Eng), M Eng, C Eng, MIET

View Full Paper

Power Analysis Based Side Channel Attacks

Abstract –Power Analysis Attacks which break the key of a by measuring the power consumption of a device have become a huge security threat. Advanced Encryption Standard (AES) which takes billions of years to break via a brute force attack can be broken in few minutes using a power analysis attack. Therefore, to minimize the threat imposed, research on countermeasures has become extremely important. First, we build a test bed for power analysis, which is a set of complete hardware and software components that can be used to do a practical demonstration of a power analysis attack. Then using the test bed, we show that even the latest encryption algorithm like Speck can be still broken in a time less than 1 hour. Despite being an add-rotate-xor cipher that does not even use substitution box operations, we show that Speck can be broken not only on an 8 bit micro controller but also on 16 bit micro controller. Next we practically test and evaluate the effectiveness of some selected countermeasures. We work on both circuit level hardware countermeasures as well as software countermeasures. While experimenting with new ideas for countermeasures finally we provide improvements to an existing countermeasure.
Keywords – Power Analysis Attack, Correlation Power Analysis, Power Measurement Test bed Speck, Countermeasures for Power Analysis Attacks

Authors- Hasindu Ganegod Ragel Department of Computer Engineering, Faculty of Engineering, University of Peradeniya, Peradeniya, Sri Lanka

View Full Paper

Fuzzy Logic Based Motion Controller For Underwater Remotely Operated Vehicle

Abstract   This work demonstrates the applicability of fuzzy controller to an Underwater Remote Operated Vehicle (ROV) for motion controlling with sensor feedback’s. Stabilizing Yaw, Pitch, Roll and depth against external disturbances are considered whilst responding to the remote maneuvering commands for forward, reverse, lateral and vertical movements. All functionalities are handled by a unified fuzzy controller. Stability and responsive of the ROV is fine-tuned by adjusting the control parameters. Performance is evaluated using  field experiments. Results show the effectiveness of fuzzy controlling of the ROV motion against external disturbances. Keywords – fuzzy controller, ROV, control theory, stabilization against disturbances, unified motion control

Authors-P.M. Saranguhewa,S.D.S. Lakmal , R.M.T. Lakmali, N.D. Jayasundere, S.H.K.K.

View Full Paper