Venue : Coral Deira - April 23 ~ 24, 2016, Dubai, UAE

Accepted Papers

  • Database Management for Human Resources Recruitment. Case of the Written
    AzeddineAmrane, University of Setif 01, Algeria
    This paper aims at summarizing how to design a database management model for the recruitment process, case of the written test, and thus illustrating the importance gained from taking the advantage of information and communication technologies.The results illustrated that the recruitment process could be automatized, where the user needs just to key in the information about candidates and marks, then everything would be sophisticated, in particular after taking on consideration SQL.
  • Multicriteria Decision Aided System for Ranking Industrial Zones (RPRO4SIGZI)
    AissaTaibi and BaghdadAtmani, University of Oran, Algeria
    Integration of Geographic Information Systems (GIS) and multi-criteria decision analysis (MCDA) is a privileged and indispensable way to evolve GIS into real decision support systems. RPRO4SIGZI, the system proposed in this paper allows, from a detailed study of geographical, environmental and socioe-conomic criteria to cooperate GIS and multi-criteria decision analysis method for spatial choosing of the right site for installing industrial projects. The result obtained by RPRO (Ranking PROMETHEE) for ranking industrial zones in western Algeria is refined by a viewing SIGZI (Geographic Information System for Industrial Zones). The RPRO unit rank industrial zones using the outranking PROMETHEE II method issue from European school and SIGZI module to the visualization of these zones on the map. RPRO4SIGZI system was designed for the evaluation of a new methodology of multi-criteria analysis guided by data mining. The objective is to show how data mining is used to model the preferences of the decision maker tainted with subjectivity and hesitance to generate suitable performance tables. Only RPRO4SIGZI system is presented in this paper.
  • A New Formalization Variation Degree-based for Gradual Patterns
    Ghallabi Fedia, The Higher Institut of Imformatics and Management in Kairouan (ISIGK), Tunisia
    In this paper we aim to present a novel formalization of gradual patterns based on the variation degree. Gradual patterns of the form “The more/less A, the more/less B” express the co-variation between the attributes A and B. In fact, in the last few years and over the last decade, in spite of more and more efficient data mining tools, many framework highlight a specific computing pattern frequency’s (i.e.; the support) approachs and formalizations for different type of data set to extract gradual patterns in a fast and reliable way. However, developing new computing tools remains a great data analysis challenge for that in the new proposed formalization we provide an increased expressiveness that allows representing interesting gradual patterns in a distinctive way. Our approach can be seen also as a unification of previous approaches for extracting gradual patterns by providing a new definition of the value of the support of a gradual pattern using both ranking correlation and variation degree.
  • Predictive Analytics in Healthcare System using Data Mining Techniques: Predicting Chronic Kidney Failure
    Basma Boukenze and Haqiq Abdelkerim, Department of Applied Mathematics and Computer Science, Morocco
    The health sector has seen a great evolution following the development of new computer technologies, and that pushed this area to produce more medical data, this gave a birth to multiple fields of research. Many efforts are provided to cope with the explosion of medical data on the one hand, and on the other hand; to obtain useful knowledge from it, and that prompted researchers to apply all the technical innovations as like big data analytics, predictive analytics, machine learning and learning algorithms, in order to extract useful knowledge and help in making decision .With the promises of predictive analytics in big data , and the use of machine learning algorithms, predicting future is no longer a difficult task, especially for medicine because predicting diseases and anticipate the cure became possible. In this paper we will present an overview on the evolution of big data in healthcare system, and we apply a learning algorithm on a set of medical data , the objective is to predict chronic kidney diseases by using Decision Tree (C4.5) algorithm.
  • Mining of Web Log Files Using Relevant Computing Techniques for Improving Future Anticipation Usage of Web Navigation
    N. Sanfia Sehnaz and I.Elizabeth Shanthi, Avinashilingam Institute for Home Science and Higher Education for women university, India.
    The Internet has evolved significantly over the past few decades. Web navigation refers to the process of navigating a network of information resources in the World Wide Web, which is organized as hypertext or hypermedia. The navigation related to web navigation usability gets solved by comparing the actual and anticipated usage patterns. The actual usage pattern removed from web server logs are periodically recorded in operational websites for handling the log data. This process is used to identify the users, user session and user task oriented transactions. The pattern can be discovered among the actual usage path by using the algorithms of data mining generally the ideal user’s interactive path models are framed by cognitive experts based on the cognition of user behavior,which is utilized to pull out the anticipated usage, that includes information about both the time required for user-oriented tasks and the mechanism to identify the user navigation problems here the usability issues get detected from the deviation of the data. Genetic algorithm is used as an optimization method as a corrective action to improve the usability.
  • Image Retrieval by Proximity Using Deep Neural Network
    Daoyuan Jia and Chunping Li, Tsinghua University, China
    Associating image content with their geolocation has long been pursued. Many solutions focuses on descriptor extrac-tion and correspondence matching to measure image simi-larity by visual features, and infer the location from similar images. Other solutions utilize the correspondence of de-scriptors to reconstruct 3D model for image location calcu-lation. However, the descriptor based solutions are usually computationally heavy and the similarity measurement is un-reliable for buildings of similar textures, which is quite com-mon if the image comes from a wide area. In this paper, we leverage a modified deep neural network to estimate images geolocation similarity rather than visual appearance similar-ity, to deal with cases where images of different location but with resembling visual features. Our model builds upon the ability of neural network to extract high level features and the back-propagation can maximize the tiny difference of vis-ually-similar images. We collected tens of thousands photos on campus, with many visually resembling, but different lo-cation ones. Experiments show that using geolocation simi-larity measurement can correctly deal with such cases.