Impact of Exponent Parameter Value for the Partition Matrix on the Performance of Fuzzy C Means Algorithm
This work addresses a parameter tuning issue for researchers using fuzzy clustering algorithms, but it is incremental as it focuses on a known parameter without introducing new methods.
The paper experimentally analyzes how the exponent parameter in the partition matrix affects the performance of the Fuzzy C Means algorithm, finding that its value significantly influences clustering outcomes, though specific numerical results are not provided.
Soft Clustering plays a very important rule on clustering real world data where a data item contributes to more than one cluster. Fuzzy logic based algorithms are always suitable for performing soft clustering tasks. Fuzzy C Means (FCM) algorithm is a very popular fuzzy logic based algorithm. In case of fuzzy logic based algorithm, the parameter like exponent for the partition matrix that we have to fix for the clustering task plays a very important rule on the performance of the algorithm. In this paper, an experimental analysis is done on FCM algorithm to observe the impact of this parameter on the performance of the algorithm.