• Login
    View Item 
    •   SU+ Home
    • Research and Publications
    • Strathmore Institute of Mathematical Sciences (SIMs)
    • SIMs Projects, Theses and Dissertations
    • BBSF Research Projects
    • BSSF Research Projects (2018)
    • View Item
    •   SU+ Home
    • Research and Publications
    • Strathmore Institute of Mathematical Sciences (SIMs)
    • SIMs Projects, Theses and Dissertations
    • BBSF Research Projects
    • BSSF Research Projects (2018)
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Fraud detection using machine leaning: a comparative analysis of neural networks & support vector machines

    Thumbnail
    View/Open
    Full-text Undergraduate project (21.61Mb)
    Date
    2018
    Author
    Gitonga, Joseph Theuri
    Metadata
    Show full item record
    Abstract
    Fraud detection and prevention tools have been evolving over the past decade with the ever growing combination of resources, tools, and applications in big data analytics. The rapid adoption of a new breed of models is offering much deeper insights into data. There are numerous machine learning techniques in use today but irrespective of the method employed the objective remains to demonstrate comparable or better recognition performance in terms of the precision and recall metrics. This study evaluates two advanced Machine Learning approaches: Support Vector Machines and Neural Networks while taking a look at Deep Learning. The aim is to identify the approach that best identifies fraud cases and discuss challenges in their implementation. The approaches were evaluated on real-life credit card transaction data. Support Vector Machines demonstrated overall better performance across the various evaluation measures although Deep Neural Networks showed impressive results with better computational efficiency.
    URI
    http://hdl.handle.net/11071/6496
    Collections
    • BSSF Research Projects (2018) [18]

    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV
     

     

    Browse

    All of SU+Communities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    Login

    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV