SiMCAL 1 algorithm for analysis of gene expression data related to the phosphatidylserine receptor

GAKREM: A novel hybrid clustering algorithm

“…Similarly, the 1st, 3rd, 5th and 8th data points are the medoids of the four clusters represented by the chromosome 3. To optimize the use of computer memory, instead of storing the full length of chromosomes, we encode only the positions of 1-loci in chromosomes, thus the chromosomes in the above example can be represented as a set of ordered numbers (4,9), (2,6,7,10), and (1, 3, 5, 8), respectively. Each chromosome now specifies the k number of clusters plus the medoids of clusters in the data.…”

“…In other words, they are very sensitive to initialization parameters and may suffer from over-fitting (too many clusters generated) while not reflecting the true underlying structure [9,16,17,22]. 0020 There are several well-known algorithms such as statistical clustering [9,33,34], hierarchical clustering [10,16,31], partitional clustering [22,35], nearest neighbor clustering [19], fuzzy C-means clustering [3,26], and neural network clustering [18]. Among them, the K-means is the most popular because of its simplicity and computational efficiency although it is very sensitive to the initial choice of medoids [17,22].…”