Multidimensional Network Analysis for Analyzing and Predicting Complex Customer-Product Relations in Engineering Design

点击翻译按钮获取中文摘要

Understanding customer preferences and needs is critically important in developing successful products. This award supports an interdisciplinary research to develop a data-driven mathematical approach for analyzing and predicting consumer preferences in design of sustainable engineering products, such as alternative fuel vehicles and smart appliances. The relations among customers and products are conceptualized as a complex network and analyzed using network theory and techniques. This study will help industry produce more competitive products in shorter time to market. The findings will contribute to the development of new techniques for analyzing large complex networks. Workshop and panel sessions on analyzing customers and products as networks will be organized for dissemination to a broader community. Research will be integrated with education through interdisciplinary undergraduate Design Certificate and graduate Design Cluster programs. Analytical modeling of customer preferences in product design is inherently difficult as it faces challenges in modeling heterogeneous human behavior and product offerings. The novelty of the research lies in the employment of a Multidimensional Customer-Product Network (MCPN) framework, where separate networks of "customers" and "products" are simultaneously modeled, and multiple types of relations, such as consideration and purchase, product associations, and customer social networks are considered. The research will extend the Exponential Random Graph Model (ERGM) as a unified statistical inference framework for analyzing multidimensional customer-product relations and predicting unknown customer preferences (consideration or choice) under new design scenarios. Social influences on adopting "green" technology are analyzed in the same framework. Our approach overcomes the limitations of the traditional statistical analysis and utility-based preference modeling by considering the dependency among product choices and the social influence induced "irrationality" of customer behavior. We will also explore the use of text analysis of customer-generated data in social media thereby creating crowdsourced "virtual labs" for advancing data analytics and computational social science in product design.