Accepted Papers

RAIN STREAKS REMOVAL FROM VIDEOS

Dinesh Kadam¹, Amol R. Madane² and S. V. Bonde¹, ¹Department of Electronics and Telecommunication, SGGSIET, Nanded, India,²Tata Consultancy Services Ltd., Pune, India.

ABSTRACT

Rain streak removal from scene is important and lot of research in this area, robust and real time algorithms are still unavailable into market. Difficulties in rain removal algorithm are arises due to less visibility, less illumination, availability of moving camera and objects. Challenges in rain streak recovery algorithms are detection of rain streaks and replace them with original values to recover the scene. In this paper, photometric and chromatic properties are used for rain detection. Moving objects have detected by updated Gaussian Mixture Model (Updated GMM). This rain streak removal algorithm is used to detect rain streaks from videos and replace it with estimated values which is equivalent to original value. The concept of Inpainting is used to replace rain streaks with its original values. The spatial and temporal properties are used.

KEYWORDS

Dynamic Scene, Edge Filters, Gaussian Mixture Model (GMM), Rain Streaks Removal, Scene Recovery, Video Deraining.

SOLVING THE BINARIZATION CHALLENGES IN DOCUMENT IMAGES USING OTSU MULTILEVEL

Enas M. Elgbbas, Mahmoud I. Khalil, Hazem Abbas, Faculty of Engineering, Ain Shams University, Cairo, Egypt

ABSTRACT

This paper introduces a method for binarization of historical document images that suffer from non-uniform background, faint text, low contrast, stain, bleed-through, shadow challenges. The proposed method adaptively detects the non-uniform background in the document image and eliminate it. Areas that contain missing text are adaptively identified and reprocessed separately. Stain and bleed-through objects are found depending on stroke width and locally binarized. Shadow is detected based on the image contrast. Otsu multilevel is applied for binarization. DIBCO series is used for testing.

KEYWORDS

Document image binarization, Otsu multilevel

TEMPORAL ANALYSIS OF ELECTROMYOGRAM SIGNALS (EMGS)

Fadia meziani and Souhila Rerbal, Genie Biomedical Laboratory (GBM),Faculty of Technology,University A.B.Belkaid-Tlemcen BP 119 Tlemcen, Algeria

ABSTRACT

The aim of this study is to analyse EMGs signals using temporal analysis, in order to provide a wide range of information’s related to the type of signal (normal and pathological)

KEYWORDS

Electomyogram EMG, Signal, Temporal Analysis, Parameters &Pathology.

The Statistical analysis of the live TV bit rate

Iskandar Aripov, Master of Science in Computer Science, Georgia Institute of Technology, Atlanta, GA, USA

ABSTRACT

Currently there is a wide range of solutions for thelive TV streaming based on statistical multiplexing techniques.This paper studies the statistical nature of TV channelsstreaming variable bit rate distribution and allocation. Thegoal of the paper is to derive the best-fit rate distribution todescribe TV streaming bandwidth allocation, which can revealtraffic demands of users and be used to improve the efficiencyof the relative networking devices. Our analysis usesmultiplexers channel bandwidth allocation (PID) data of 13 TVlive channels monitored during 23 minutes and covers 36673measurements (2821 per channel). We use MATLAB and asupplementary script [1] to apply 17 continuous and 3 discretedistributions to determine the best-fit distribution function foreach individual channel and for the whole set of channels. Wefound that the “generalized extreme value” and “generalizedpareto” distribution functions are fitting most of our channelsmost precisely according to the Bayesian information criterion.By the same criterion “tlocationscale” distribution matchesbest for the whole system. We use these results to proposeparameters for streaming server queuing model. Results areuseful for streaming servers scheduling policy design processtargeting to improve limited infrastructural resources, trafficengineering through dynamic routing at CDN, SDN.

KEYWORDS

Statistical Multiplexing, Video compression &Streaming, Video Analysis and Event Recognition, InternetSignal Processing, Streaming bit rate, bandwidth utilization, TVchannels, SD quality, distribution function, probability,Communication theory and techniques, random signals, queuingtheory, Hurst exponent.

CHEMCONNECT: AN ONTOLOGY-BASED REPOSITORY OF EXPERIMENTAL DEVICES AND OBSERVATIONS

Edward S. Blurock, Blurock Consulting AB, Sweden.

ABSTRACT

CHEMCONNECT is an ontology cloud-based repository of experimental, theoretical and computational data for the experimental sciences domain. Currently, the emphasis is on the chemical combustion community, but future work (in collaboration with domain experts), the domain will be expanded. CHEMCONNECT goes beyond traditional meta-data annotated scientific result repositories in that the data is parsed and analysed with respect to an extensive chemical and combustion knowledge base. The parsed data is then inter-linked allowing for efficient searching and comparison. The goal is to have all data associated with experiments, from a device description, to the intermediate data (both computed and measured) and to their associated interpretations and the procedures and methodologies to the final published results and references to be available. Having published data linked to its dependent measurements and constants, devices, subsystems, sensors and even people and laboratories provides an effective accountability and more confidence in the data. Data entry and availability can range from private user, to user defined consortia to general public. These concepts were implemented at http://www.connectedsmartdata.info.

KEYWORDS

Case Study, Ontology, Repository, Database, Experimental Devices, Experimental Results.

Resolution Enhancement of Electron Microscopic Volume by Volume Restoration Technique

Anik Khan¹, Kishor Datta Gupta¹, and Ariful Haque², ¹Department of Computer Science, University Of Memphis , Memphis,TN, USA,3811 and ²Department of Electrical and Electronic Engineering , Bangladesh University of Engineering and Technology , Dhaka-1205

ABSTRACT

The knowledge of structure of proteins, protein derived compounds and RNA structures in eukaryotic cell is mandatory to understand the functions of these macromolecules.With recent development of Direct Electron Detector Device (DDD) camera and application of maximum likelihood algorithms in volume reconstruction, cryo-Electron Microscopy (cryo-EM) enables us to visualize the macromolecules in nearly physiological state. The current resolution limit of cryo-EM can be improved further by applying novel and effective signal processing algorithms after the EM workflow. In this work, a signal processing method has been developed to enhance the resolution of the EM volume through volume restoration techniques. We have proposed a novel technique to estimate the volume degradation function of the volume reconstruction system from the noise-only subvolumes of the observed EM volume. Then the volume is restored (utilizing the estimated volume degradation function) using a combination of regularized Richardson-Lucy and Wiener Deconvolution techniques. In addition to volume restoration, we have employed spatial de-noising techniques utilizing various morphological filters to reduce noise outside the main molecular structure. The experimental results demonstrate that the resolution (evaluated by Fourier Shell Correlation curve) and visual quality of the EM volume can be significantly improved using our proposed technique.

KEYWORDS

Electron Microscopic, Signal processing, Visualization, volume degradation function, Image processing

CADS: CORE-AWARE DYNAMIC SCHEDULER FOR MULTICORE MEMORY CONTROLLERS

Eduardo Olmedo Sanchez, Technical university of Madrid Calle Jose Gutiérrez Abascal, 2 28006 Madrid

ABSTRACT

Memory controller scheduling is crucial in multicore processors, where DRAM bandwidth is shared. Since increased number of requests from multiple cores of processors becomes a source of bottleneck, scheduling the requests efficiently is necessary to utilize all the computing power these processors offer. However, current multicore processors are using traditional memory controllers, which are designed for single-core processors. They are unable to adapt to changing characteristics of memory workloads that run simultaneously on multiple cores. Existing schedulers may disrupt locality and bank parallelism among data requests coming from different cores. Hence, novel memory controllers that consider and adapt to the memory access characteristics, and share memory resources efficiently and fairly are necessary. We introduce Core-Aware Dynamic Scheduler (CADS) for multicore memory controller. CADS uses Reinforcement Learning (RL) to alter its scheduling strategy dynamically at runtime. Our scheduler utilizes locality among data requests from multiple cores and exploits parallelism in accessing multiple banks of DRAM. CADS is also able to share the DRAM while guaranteeing fairness to all cores accessing memory. Using CADS policy, we achieve 20% better cycles per instruction (CPI) in running memory intensive and compute intensive PARSEC parallel benchmarks simultaneously, and 16% better CPI with SPEC 2006 benchmarks.

KEYWORDS

multicore processors, reinforcement learning, high performance computing, memory controller, machine learning

A MACHINE LEARNING ALGORITHM IN AUTOMATED TEXT CATEGORIZATION OF LEGACY ARCHIVES

Dali Wang¹, Ying Bai² and David Hamblin³, ¹Christopher Newport University, Newport News, VA, USA,²Johnson C. Smith University, Charlotte, NC, USA, ³Christopher Newport University, Newport News, VA, USA.

ABSTRACT

The goal of this research is to develop an algorithm to automatically retrieve critical information from raw data files in NASA’s airborne measurement data archive. The product has to meet specific metrics in term of accuracy, robustness and usability, as the initial decision-tree based development has shown limited applicability due to its resource intensive characteristics. We have developed an innovative solution that is much less resource intensive while offering comparable performance. As with many practical applications, the data available are noisy and correlated; and there is a wide range of features that are associated with the information to be retrieved. The proposed algorithm uses a decision tree to select features and determine their weights. A weighted Naive Bayes is used due to the presence of highly correlated inputs. The development has been successfully deployed in an industrial scale, and the results show that the development is well-balanced in term of performance and resource requirements.

KEYWORDS

Machine Learning, Classification, Naïve Bayes, Decision Tree.

PREDICTING CUSTOMER CALL INTENT BY ANALYZING PHONE CALL TRANSCRIPTS BASED ON CNN FOR MULTI-CLASS CLASSIFICATION

Junmei Zhong and William Li, Marchex Inc 520 Pike Street, Seattle, WA, USA.

ABSTRACT

Auto dealerships receive thousands of calls daily from customers interested in sales, service, vendors and jobseekers. With so many calls, it is very important for auto dealers to understand the intent of these calls to provide positive customer experiences that ensure customer satisfaction, deeper customer engagement to boost sales and revenue, and optimum allocation of agents or customer service representatives across the business. In this paper, we define the problem of customer phone call intent as a multi-class classification problem stemming from the large database of recorded phone call transcripts. To solve this problem, we develop a convolutional neural network (CNN)-based supervised learning model to classify the customer calls into four intent categories: sales, service, vendor or jobseeker. Experimental results show that with the thrust of our scalable data labeling method to provide sufficient training data, the CNN-based predictive model performs very well on long text classification according to tests that measure the model’s quantitative metrics of F1-Score, precision, recall, and accuracy.

KEYWORDS

Word Embeddings, Machine Learning, Deep Learning, Convolutional Neural Networks, Artificial Intelligence, Auto Dealership Industry, Customer Call Intent Prediction.

AHP UNDER UNCERTAINTY: A MODIFIED VERSION OF CLOUD DELPHI HIERARCHICAL ANALYSIS

Alaa Abou Ahmad¹, Ghaida Rebdawi² and Obaida alsahli³,¹ Information Department, HIAST, Damascus, Syri,² Information Department, HIAST, Damascus, Syria,³ Quality Program, SVU, Damascus, Syria.

ABSTRACT

Cloud Delphi Hierarchical Analysis (CDHA) is an Analytic Hierarchical Process (AHP) based method for group decision making under uncertain environments. CDHA adopts appropriate tools for such environments, namely Delphi method, and Cloud model. Adopting such tools makes it a promising AHP variant in handling uncertainty. In spite of CDHA is a promising method, it is still suffering from two main defects. The first one lies in its definition of the consistency index, the second one lies in thetechnique used in building the pairwise comparisons Cloud models. This paper will discuss these defects, and propose a modified version. To overcome the defects mentioned above, the modified version will depend more on the context of the interval pairwise comparisons matrix while building the corresponding Cloud pairwise comparisons matrix. A simple case study that involves reproducing the relative area sizes of four provinces in Syria will be used to illustrate the modified version and to compare it with the original one.

KEYWORDS

Fuzzy Set Theory Based AHP, Multi-criterion Group Decision Making, Decision Making under Uncertainty, Cloud Model, Delphi Method.

EVOLVING RANDOM TOPOLOGIES OF SPIKING NEURAL NETWORKS FOR PATTERN RECOGNITION

Gustavo López-Vázquez¹, Manuel Ornelas-Rodríguez¹, Andrés Espinal², Jorge A.Soria-Alcaraz², Alfonso Rojas-Domínguez¹, Héctor J. Puga-Soberanes¹, J.Martín Carpio¹and Horacio Rostro-González², ¹Division of Postgraduate Studies and Research, National Technology of México /León Institute of Technology. León, Guanajuato, México ²Department of Organizational Studies (DCEA), University of Guanajuato.Guanajuato, Guanajuato, México.

ABSTRACT

Artificial Neural Networks (ANNs) have been successfully used in Pattern Recognition tasks. Evolutionary Spiking Neural Networks (ESNNs) constitute an approach to design third-generation ANNs (also known as Spiking Neural Networks, SNNs) involving Evolutionary Algorithms (EAs) to govern some intrinsic aspects of the networks, such as topology, connections and/or parameters. Concerning the practicality of the networks, a rather simple standard is commonly used; restricted feed-forward fullyconnected network topologies deprived from more complex connections are usually considered. Notwithstanding, a wider prospect of configurations in contrast to standard network topologies is available for research. In this paper, ESNNs are evolved to solve pattern classification tasks, using a EAbased algorithm known as Grammatical Evolution (GE) which enables a higher degree of freedom in automatically determining partially-connected networks, their topologies (number of layers, units and synapses), and their connections types (looped, unrestricted, supralayer). Experiments show competitive results and a distinctive variety of network designs when compared to a more traditional approach to design ESNNs.

KEYWORDS

Artificial Neural Networks, Spiking Neural Networks, Evolutionary Spiking Neural Networks,Evolutionary Algorithms, Grammatical Evolution

An Enhanced Ad Event-Prediction Method Based on Feature Engineering

Saeid Soheily-Khah and Yiming Wu, SKYLADS Research Team, Paris, France

ABSTRACT

In digital advertising, Click-Through Rate (CTR) and Conversion Rate (CVR) are very important metrics for evaluating ad performance. As a result, ad event prediction systems are vital and widely used for sponsored search and display advertising as well as Real-Time Bidding (RTB). In this work, we introduce an enhanced method for ad event prediction (i.e. clicks, conversions) by proposing a new efficient feature engineering approach. A large real-world event-based dataset of a running marketing campaign is used to evaluate the efficiency of the proposed prediction algorithm. The results illustrate the benefits of the proposed approach, which significantly outperforms the alternative ones.

KEYWORDS

Digital Advertising, Ad Event Prediction, Feature Engineering, Feature Selection, Classification

WEB BASED COMPUTER TROUBLESHOOTING EXPERT SYSTEM

Henok Yared Agizew and Askale Tesfaye Aga, Mettu University, Mettu, Ethiopia

ABSTRACT

Computer has two main parts or components such as hardware and software. Each computer parts may faced with several troubles that makes those computers out of services. In order to overcome such problems with those computer parts, it requires ICT proffesionals with appropriate resources of troubleshooting techniques. Inversily, only limmited number of profissionals and resources which are used to troubleshooting the problem and provide the solution easly are there. An expert system is a part of artificial intelligent application that provide solution for complicated problems that would otherwise require extensive human expertise. The aim of this study is developing a prototype Web-based expert system whith the ability and capability to delivering appropriate computer troubleshooting advices for the information and communication technology technicians and other users based on the problem occurred in the computer system. Once the domain knowledge is collected, modeled and represented using semi-structured interview technique, decision table and ‘if–then’ rules techniques respectively, the web-based expert system was developed by using e2gLite expert system develpment shell.

KEYWORDS

Expert System, Computer Troubleshooting, Web based system

Designing Neuro-Classifier Fusion by Extreme Learning Machine and NSGA-II: A New Direction for Building Machine-Readable Arabic Offline Character Recognition

Saad M. Darwish, Khaled O. Elzoghaly

Institute of Graduate Studies and Research, Alexandria University 163 Horreya Avenue, El Shatby 21526. P.O. Box 832, Alexandria, Egypt.

ABSTRACT

In recent years, there was intensive research on Arabic Optical Character Recognition (OCR) especially the recognition of scanned, offline, machine-printed documents. However, Arabic OCR results are unsatisfactory and are still an evolving research area. Exploring the best feature extraction techniques and selecting an appropriate classification algorithm lead to superior recognition accuracy and low computational overhead. This paper presentsa new Arabic OCR approach by integrating both of Extreme Learning Machine (ELM) and Non-dominated Sorting Genetic Algorithm (NSGA-II) in a unified framework with the aim of enhancing recognition accuracy. ELM is adopted as a neural network classifier that has a short processing time and avoids many difficulties faced by gradient-based learning methods such as learning epochs and local minima. NSGA-II is utilized as a feature selection algorithm that has better convergence and spread of solutions. NSGA-II emphasizes non-dominated solutions and explicit diversity preservation mechanism. Experimental results compared to other approaches revealed the efficiency of the proposed system and demonstrated that the feature selection approach increased the accuracy of the recognition process.

KEYWORDS

Index Terms—Arabic OCR, Extreme Learning Machine, Feature selection, NSGA-II.

TOWARDS AN UNDERSTANDING OF THE DIFFERENCE BETWEEN SOFTWARE QUALITY CHARACTERISTICS AND SOFTWARE SUSTAINABILITY CHARACTERISTICS

Sulaiman Aljarallah and Russell Lock, Department of computer science, Loughborough University, Loughborough, UK.

ABSTRACT

Software sustainability has generated much interest in the software engineering field in recent times. The interference between software quality and software sustainability is still an open question. In this study, the comparison was conducted using a three-phases, having as a starting point the comparison of basic quality models to determine the most frequent characteristics between these models. A follow-up study, conducted at a more comprehensive level to cover both basic models and the most cited tailored models, shows nearly the same results as the comparison of basic models. Therefore, both studies show nearly the same characteristics. Software sustainability literature is investigated to find the most frequent characteristics. The study concludes that the most frequent characteristics are maintainability, portability, usability and efficiency, followed by reliability, reusability, security, durability and extensibility. The study suggests that ISO 25010 can be utilised by software sustainability.

KEYWORDS

Software sustainability, Software quality, Software sustainability characteristics, Quality characteristics. Quality models.

New Meta characters to analyse data using regular expressions

Boulchahoub Hassan, Labriji amine, Labriji El Houssine, Rachiq Zineb, Gourram Hicham, Mohamed AZOUAZI, Department of Mathematics and Computer Science Faculty of Sciences Ben M’SIK Casablanca, Morocco

ABSTRACT

Regular expressions are often used to search for elements with well-defined structures such as emails, phone numbers or Ip addresses... They also make it possible to format the found elements and to change their way of appearance in strings or files. To treat all the possible cases, and to obtain precise search results, several metacharacters have been reserved and integrated into the dictionary of regular expressions including \, ^, $, {,}, [,], (,),., *, +, |?, <,>, -, &, etc. Each Meta-character has a precise mission and an exact role when parsing Strings. Unfortunately these metacharacters do not allow to analyze the results found by a regular expression, they do not allow to make some restrictions such as (equal, not equal, between) or some projections like (maximum, average, grouping by ..) or sorts. Currently, to do these treatments, we must implement our own algorithms which cost a remarkable effort and a waste of time. We propose in this paper to add some new meta characters to analyze the results of a regular expression. New meta-characters for making restrictions, projections and sorting automatically.

KEYWORDS

Regular expression, Meta character, big data, finite automata

COMPARISON OF SCRUM AND KANBAN IN THE LEARNING MANAGEMENT IMPLEMENTATION PROCESS

Aida Granulo¹ and Anel Tanovic²

¹Department of Information Technologies, International Burch University, Sarajevo, Bosnia and Herzegovina, ²Department of Computer Science, Faculty of Electrical Engineering Unviersity OF Sarajevo, Sarajevo, Bosnia and Herzegovina

ABSTRACT

In this paper two methods of SCRUM and Kanban are compared to the example of building a Learning Management System. A comparative analysis of SCRUM as a method based on planning, people and a clear organization, and Kanban, which puts its focus on flexibility, tasks and processes, delivery and visualization of the processes, has been made. The calculations specific to each of the technologies give a clear view of the time needed to complete the project and the amount of work that needs to be done if the project duration is clearly limited. When time is not a limiting factor, Kanban gave better results and the project would end two months earlier than when the Scrum methodology was applied. On the other hand, when it is necessary to carry out detailed monitoring and observe time limit Scrum has yielded slightly better results. Based on these results, guidelines that can help in making decision of the method to use for a specific project are given.

KEYWORDS

SCRUM, KANBAN, agile software methodology, Learning Management System, velocity, WIP

REVIVING LEGACY ENTERPRISE SYSTEMS WITH MICROSERVICE-BASED ARCHITECTURE WITHIN CLOUD ENVIRONMENTS

Safa Habibullah, Xiaodong Liu and Zhiyuan Tan, School of Computing, Edinburgh Napier University, Edinburgh, UK and Zhang, Qi Liu, School of Automation, Nanjing University of Information Science and Technology, China

ABSTRACT

Evolution has always been a challenge for enterprise computing systems. The microservice based architecture is a new design model which is rapidly becoming one of the most effective means to re-architect legacy enterprise systems and to reengineer them into new modern systems at a relatively low cost. This architectural style has evolved based on a number of different approaches and standards. However, there are quite a few technical challenges which emerge when adopting microservices to revive a legacy enterprise system. In this paper, an evolution framework and a set of feature-driven microservices-oriented evolution rules have been proposed and applied to modernise legacy enterprise systems, with a special emphasis on analysing the implications as regards runtime performance, scalability, maintainability and testability. Testing and evaluation have been carried out in depth, aiming to provide a guidance for the evolution of legacy enterprise systems.

KEYWORDS

Microservice, Legacy System, Software Evolution, Cloud Environment

AN APPROACH TO TRACKING PROBLEM FOR LINEAR CONTROL SYSTEM VIA INVARIANT ELLIPSOIDS METHOD

Mikhail Khlebnikov, Laboratory of Adaptive and Robust Systems, V. A. Trapeznikov Institute of Control Sciences of Russian Academy of Sciences,Moscow, Russia.

ABSTRACT

In this paper, a simple yet universal approach to the tracking problem for linear control systems via the linear static combined feedback is proposed. The approach is based on the invariant ellipsoid concept and LMI technique, where the optimal control design reduced to finding the minimal invariant ellipsoid for the closedloop system. With such an ideology, the control design problem directly reduces to a semidefinite programming and one-dimensional minimization. Another attractive property of the proposed approach is that it is equally applicable to discrete-time control systems. The efficacy of the technique is illustrated via a benchmark problem.

KEYWORDS

Linear Control Systems, Tracking Problem, Invariant Ellipsoids, LMIs.

HMM-Based Dari Named Entity Recognition for Information Extraction

Ghezal Ahmad Jan Zia¹ and Ahmad Zia Sharifi², ¹Technical University of Berlin, Germany and ²Nangarhar University, Afghanistan.

ABSTRACT

Named Entity Recognition (NER) is the fundamental subtask of information extraction systems that labels elements into categories such as persons, organizations or locations. The task of NER is to detect and classify words that are parts of sentences. This paper describes a statistical approach to modeling NER on the Dari language. Dari and Pashto are low resources languages, spoken as official languages in Afghanistan. Unlike other languages, named entity detection approaches differ in Dari. Since in Dari language there is no capitalization for identifying named entities. We seek to bridge the gap between Dari linguistic structure and supervised learning model that predict the sequences of words paired with a sequence of tags as outputs. Dari corpus was developed from the collection of news, reports and articles based on the original orthographic structure of the Dari language. The experimental result presents the named entity recognition performance 95% accuracy.

KEYWORDS

Natural Language Processing (NLP), Hidden Markov Model(HMM), Named Entity Recognition (NER), Part-of-Speech (POS) Tagging.

Video Processing and analysis by SVM on SPARK

Yiyang Teng and Bin Wu, Beijing Key Lab of Intelligent Telecommunication Software and Multimedia, China.

ABSTRACT

Currently, with the explosion of video data, analysis and processing of large-scale video data has gradually emerged. At the same time, parallel framework on Spark has shown unprecedented advantages in machine learning tasks. However, the Spark framework is mostly used to process text data, and is less used to process other data such as video. In this paper, we propose a large-scale video processing framework with parallel SVM based on Spark. By serializing video frames and introducing OpenCV, CaffeOnSpark and other open source tools, we realize the image feature extraction and face/expression recognition of large-scale video data combine with MSP-SVM algorithm. At the same time, in order to make the calculation process more efficient, we propose two data load balancing methods to accelerate parallel computing. Experimental results show that our parallel framework is effective for video processing on the large-scale data set.

KEYWORDS

Support Vector Machine, Spark, Parallel Computing, Face Recognition, Expression Classification.

FACTORS AFFECTING CLASSIFICATION ALGORITHMS RECOMMENDATION: A SURVEY

Mariam Moustafa Reda, Dr. Mohammad Nassef and Dr. Akram Salah, Computer Science Department, Faculty of Computers and Information, Cairo University, Giza, Egypt.

ABSTRACT

A lot of classification algorithms are available in the area of data mining for solving the same kind of problem with a little guidance for recommending the most appropriate algorithm to use which gives best results for the dataset at hand. As a way of optimizing the chances of recommending the most appropriate classification algorithm for a dataset, this paper focuses on the dif erent factors considered by data miners and researchers in dif erent studies when selecting the classification algorithms that will yield desired knowledge for the dataset at hand. The paper divided the factors af ecting classification algorithms recommendation into business and technical factors. The technical factors proposed are measurable and can be exploited by recommendation software tools.

KEYWORDS

Classification, Algorithm selection, Factors, Meta-learning, Landmarking

A Comparative Mention-Pair Models for Coreference Resolution in Dari Language for Information Extraction

Ghezal Ahmad Jan Zia¹, Ahmad Zia Sharifi², Fazl Ahmad Amini³, and Niaz Mohammad Ramaki⁴ , ¹Technical University of Berlin, Berlin Germany , ²Nangarhar University, Nangarhar, Afghanistan , ³Faculty of Literature, Kabul University , ⁴Kabul Polytechnique University, Kabul, Afghanistan

ABSTRACT

Coreference resolution plays an important role in Information Extraction.This paper covers the investigation of two strategies based on a mention-pair resolver using Decision Tree classifier on structured and unstructured dataset, targeting coreference resolution in Dari language. Strategies are (1) training separate models which is specialized in particular categories(e.g., lexical, syntactic and semantic) and types of mentions (e.g. pronouns, proper nouns) and (2) using a structured dataset on a machine learning library that is designed to classify numerical values. Moreover, these modifications and comparative models describe a contribution of comprehensive factors involved in the resolution of texts. Specifically, we developed the first Dari corpus(’DariCoref’) based on OntoNotes and WikiCoref scheme. Both strategies are produced f-score of state-of-the-art.

Vulnerability Analysis of IP Cameras Using ARP Poisoning

Thomas Doughty¹ and Usman Adeel², ¹BSc (Hons) Cyber Security and Networks and ²Senior Lecturer in Computer Science, Teesside University, Middlesbrough

ABSTRACT

Internet Protocol (IP) cameras and Internet of Things (IoT) devices are known for their vulnerabilities, and Man in the Middle attacks present a significant privacy and security concern. Because the attacks are easy to perform and highly effective, this allows attackers to steal information and disrupt access to services. We evaluate the security of six IP cameras by performing and outlining various attacks which can be used by criminals. A threat scenario is used to describe how a criminal may attack cameras before and during a burglary. Our findings show that IP cameras remain vulnerable to ARP Poisoning or Spoofing, and while some cameras use Digest Authentication to obfuscate passwords, some vendors and applications remain insecure. We suggest methods to prevent ARP Poisoning, and reiterate the need for good password policy.

KEYWORDS

Security, Camera, Internet of Things, Passwords, Sniffing, Authentication

Genetic Algorithm Based User Pairing For Noma Downlink Systems

Kaan Zorluer^1,2 and Ömer Faruk Gemici², ¹Yıldırım Beyazıt University, Turkey and ²TÜBİTAK BİLGEM, Gebze, Kocaeli, Turkey

ABSTRACT

Non-Orthogonal Multiple Access (NOMA) as a multiple access scheme can be instrumental to satisfy the high system capacity requirement of the 5G. NOMA increases the spectral efficiency by providing simultaneous transmission of multiple users at the same radio resource at the transmitter and employing more sophisticated signal processing techniques at the receiver such as successive interference cancellation (SIC). The user group selection is one of the important elements that affects the performance of NOMA. In this paper, the Genetic Algorithm approach is proposed to determine the user pair selection for multi-user NOMA downlink system. The proposed GA scheme can yield a reasonable solution for any underlying objective as long as the fitness function is properly set. The simulation results demonstrate that the GA approach is a powerful meta-heuristic to reach target objectives. However, a slight performance degradation in the GA approach is observed when the search space is relatively high.

KEYWORDS

Mobile Network, NOMA, Genetic Algorithm, Resource Allocation