Data Science and Digital Collaborations

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 4 of 4
  • Item
    I Understand What You Are Saying: Leveraging Deep Learning Techniques for Aspect Based Sentiment Analysis
    ( 2019-01-08) Tao, Jie ; Zhou, Lina ; Feeney, Conor
    Despite widespread use of online reviews in consumer purchase decision making, the potential value of online reviews in facilitating digital collaboration among product/service providers, consumers, and online retailers remains under explored. One of the significant barriers to realizing the above potential lies in the difficulty of understanding online reviews due to their sheer volume and free-text form. To promote digital collaborations, we investigate aspect based sentiment dynamics of online reviews by proposing a semi-supervised, deep learning facilitated analytical pipeline. This method leverages deep learning techniques for text representation and classification. Additionally, building on previous studies that address aspect extraction and sentiment identification in isolation, we address both aspects and sentiments analyses simultaneously. Further, this study presents a novel perspective to understanding the dynamics of aspect based sentiments by analyzing aspect based sentiment in time series. The findings of this study have significant implications with regards to digital collaborations among consumers, product/service providers and other stakeholders of online reviews.
  • Item
    Developing a User Typology for the Analysis of Participation in Enterprise Collaboration Systems
    ( 2019-01-08) Schwade, Florian ; Schubert, Petra
    In this paper, we propose a user typology for Enterprise Collaboration Systems (ECS). We draw on and extend findings from previous research in the area of CSCW and Social Collaboration Analytics. The proposed typology includes: (1) a definition of user types, (2) dimensions of ECS use and (3) a classification of action (event) types. The typology contains the following user types: creator, contributor, lurker, inactive and non-user. These types are characterized by differences in the following dimensions: type of use, frequency of use, variety of use, choice of content type and platform preferences. The definition of user types along these dimensions facilitates the implementation of database queries (scripts) for Social Collaboration Analytics (SCA), with the aim of determining the dis-tribution of types of users in an Enterprise Collaboration System. We present selected results of such SCA for an integrated collaboration platform and discuss the findings. We successfully demonstrate that our classification of user types allows us to draw conclusions on (1) the form and degree of participation of users in the ECS and, derived from that, (2) the likely purpose of the examined communities.
  • Item
    Divergence Based Non-Negative Matrix Factorization for top-N Recommendations
    ( 2019-01-08) Haque, Md. Enamul ; Zobaed, SM ; Tozal, Mehmet Engin ; Raghavan, Vijay
  • Item
    Introduction to the Minitrack on Data Science and Digital Collaborations
    ( 2019-01-08) Iyer, Lakshmi ; Paul, Souren ; Zhou, Lina