Access Type

Open Access Dissertation

Date of Award

January 2015

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Industrial and Manufacturing Engineering

First Advisor

Kyoug-yun Kim

Abstract

Crowdsourcing has emerged as a new design resource for conceptual design process and multiple crowdsourcing services provide an opportunity for design idea collection and concept generation by crowds. However, few formal methods are available to extract and evaluate design concepts from the activities of the design crowd. Scarcity of information and non-guaranteed quality of contributions are often challenges to be tackled. To overcome the challenges, the research aims to answer how a system systematically extracts and represents the explicit or implicit hidden design concepts from crowdsourcing design activities and how crowdsourcing design activities of participants are captured as design information to develop a product in crowdsourcing platform in the perspectives of process and elements.

This research provides taxonomy of design features to represent crowdsourcing design activities. With the taxonomy, a formal concept analysis method, Galois lattices, is applied to evaluate activities of design crowd and to extract possible design concepts. Using this approach, the crowd activities are represented with design features and participant information and it allows modeling the potential design concepts with the contributions of participants. Two participant evaluating measures, Participant Individual Score and Participant Group Score, are proposed to enhance the extracted design concepts with participants' information. By employing the proposed scores and design features, this research figure out the significance of participants' behavior in crowdsourcing design. In addition, a formal method to represent the processes and elements in crowdsourcing design activities with the theory adopted from social science, Actor Network Theory. The presented method and metrics are validated with a real design data collected from a crowdsourcing service by focus group interview and precision and recall tests.

Share

COinS