The potential of using fuzzy clustering in generating consumer profile

摘要

A good consumer profile can be developed by forming a suitable group of criteria. However, the most common existing consumer profile was created based on statistical averages where overlapping between characteristics does not exist. This paper discussed the potential of using the fuzzy clustering method to develop a suitable consumer profile. Fuzzy C-Means (FCM) algorithm has been selected to construct consumer profile from two different data sets; domestic water consumption and electricity consumption data sets. The result obtained using FCM algorithm was compared to the results produced by hard clustering algorithm and statistical method. The findings showed that fuzzy clustering has the potential to create a very informative consumer profile.