The optimal choice of fuzziness parameter in load profiling: article

Load profiling is an estimated load shape for a group of customers, which can be developed from historical or current day data. Many techniques for load profiling have been reported in the past. The techniques range from conventional method to more sophisticated ones such as fuzzy c-means algorithm. This thesis presents a technique for determining the weighting exponent, m, parameter in a fuzzy c-means algorithm using fuzzy clustering validity indexes. The objective of using cluster validity index is to find the best clustering schemes. In this thesis, two fuzzy clustering validity indexes used are Non-fuzzy Index (NFI) and Xie­Beni Index (XB). It will give information on the customers cluster pattern in load profile.