CLUSTERING GENE EXPRESSION DATA WITH MIN-MAX-MEDIAN INITIALIZED FUZZY C-MEANS ALGORITHMS