Читать книгу Data Analytics in Bioinformatics - Группа авторов - Страница 42

Rather than considering and calculating medoids for the entire dataset CLARA considers a subset from the dataset with a constant size upon which PAM(partitioning around medoids) algorithm is applied. PAM algorithm aims at reducing the degree of similarity by calculating the distance between the cluster medoids and the elements in the subset space. As a result an optimal set of medoids are obtained which requires less store space when compared to other algorithms. In contrast to other partitioning algorithms this measures the degree of dissimilarity amongst the elements and the medoid of the cluster. This algorithm is applied on huge datasets producing accurate results with less computation time [17].

Drawback—Cannot deal with higher dimensional data and closely connected clusters which leaves this algorithm a second choice in gene expression data.