Λογότυπο αποθετηρίου

Πλοήγηση ανά Συγγραφέα "Paouris, George"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω
Φιλτράρετε τα αποτελέσματα πληκτρολογώντας τα πρώτα γράμματα
Τώρα δείχνει 1 - 1 από 1
  • Μικρογραφία εικόνας
    Τεκμήριο
    Allowing privacy in Distributed Machine Learning Algorithms using Spark
    Paouris, George; Athens University of Economics and Business, Department of Informatics; Vassalos, Vasilios
    The need to handle and process large scale datasets has become apparent in the recent years. As the technology grows, more and more applications generate massive and complex information whether it is from social networks like the Facebook or banks that handle millions of customers. Companies have the need to process and convert massive amounts of data to knowledge and exploit them in order to have a greater perspective of things. Thus, processing such information with machine learning algorithms, demands bigger computing power because of the high intensive procedure. Data mining algorithms in a distributed environment offers great advantages when it comes down to datasets that are evolving in complexity and size. In this thesis, we implement the C4.5 and the DBSCAN algorithms in a distributed environment using various data transfer restrictions. For our implementation, we used the open-source Apache’s Spark framework. We also used a 36-node cluster and real datasets to evaluate the proper function of the algorithms and of our methods.