A study on hybridization of fuzzy possibilistic C-means clustering and Granular Computing

1. This thesis proposed an algorithm for fuzzy possibilistic C-means clustering, based on GrC (GrFPCM). The proposed algorithm utilizes the advantages of both the FPCM algorithm and GrC, to remove noisy, cope with the uncertainty factors and alleviate the negative impact of the high dimensionality of problems and increases the quality of the clustering. The DNA microarray problem was presented, as an application of the GrFPCM algorithm. The results demonstrated that GrFPCM achieves better results than some other existing clustering methods.
2. This thesis proposed an algorithm for interval type-2 fuzzy possibilistic C-means clustering based on GGF (GIT2FPCM). This method reduces the noise factor and uncertainty of the data, and thereby increases the quality of the clustering. Furthermore, the clustering execution time decreases significantly, because of the reduced dataset size.
3. This thesis proposed a method of combining the GIT2FPCM algorithm with the PSO algorithm, namely PGIT2FPCM, to optimize the objective function and improve the quality of clustering based on the advantages of PSO and granular space.
4. This thesis proposed a method of interval type-2 fuzzy possibilistic C-means clustering based on advanced GrC, namely AGrIT2FPCM. This algorithm determines the centroids of the granules, that are created by the GrFPCM method, to improve the measurement of the distance between a granule and a centroid of the cluster. This algorithm also utilizes the advantages of IT2FPCM in processing uncertainty and noisy datasets.