Analysis of user-generated multimedia data on medication management and consumption behavior using data mining techniques

Abstract

Technology-enabled tools have been suggested as a solution to assist older adults in the management and consumption of medications. However, existing systems and studies are often limited by incomplete understanding of the potential users' behaviors. This study uses a web-based survey and photo submission system to collect and analyze user profiles and behavioral characteristics. Various data mining techniques, including association rules, clustering and classification, are used on quantified data to find important behavioral patterns, group users with similar characteristics, and discern factors related to risky medication management behaviors. This paper presents the process and results of analysis, including a detailed description of coding scheme and model development. Practical and methodological implications are also discussed. © 2013 Springer-Verlag Berlin Heidelberg.