MURAL - Maynooth University Research Archive Library



    Analysing and Predicting the Runtime of Social Graphs


    Maher, Rana and Malone, David (2016) Analysing and Predicting the Runtime of Social Graphs. In: 2016 IEEE International Conferences on Big Data and Cloud Computing (BDCloud), Social Computing and Networking (SocialCom), Sustainable Computing and Communications (SustainCom) (BDCloud-SocialCom-SustainCom), 8-10 Oct. 2016, Atlanta, GA, USA.

    [img]
    Preview
    Download (713kB) | Preview


    Share your research

    Twitter Facebook LinkedIn GooglePlus Email more...



    Add this article to your Mendeley library


    Abstract

    The explosion of Social Network Analysis (SNA) in many different areas and the growing need for powerful data analysis has emphasized the importance of in-memory big data processing in computer systems. Particularly, large-scale graphs are gaining much more attention due to their wide range of application. This rise, accompanied by a massive number of vertices and edges, led computations to become increasingly expensive and time consuming. That is why there is a move towards distributed systems or Big Data cluster(s) to provide the required computational power and memory to handle such demand of huge graphs. Thus, figuring out whether a new social graph dataset can be processed successfully on a personal machine or there is a need for a distributed system or big- memory machine is still a remaining open question. In this paper, we try to address this question by providing a comparative analysis for the performance of two of the most well known SNA tools for performing commonly used graph algorithms such as counting Triads, calculating Degree Distribution and finding Clusters which can give an indication of the possibility of carrying out the work on a personal machine. Based on these measurements, we train different supervised machine learning models for predicting the execution time of these algorithms. We compare the accuracy of the different machine learning models and provided the details of the most accurate model that can be exploited by end users to better estimate the execution time expected for processing new social graphs on a personal machine.

    Item Type: Conference or Workshop Item (Paper)
    Keywords: Social Graphs; Graph Analytics; Social Network Analysis; Graph Algorithms; Performance; Predictive Modeling;
    Academic Unit: Faculty of Science and Engineering > Mathematics and Statistics
    Item ID: 15461
    Identification Number: https://doi.org/10.1109/BDCloud-SocialCom-SustainCom.2016.62
    Depositing User: Dr. David Malone
    Date Deposited: 09 Feb 2022 10:51
    Refereed: Yes
    URI:

    Repository Staff Only(login required)

    View Item Item control page

    Downloads

    Downloads per month over past year