Search

The support vector machines (SVMs) is one of the machine learning algorithms with high classification accuracy. However, the support vector machine algorithm has a very high training complexity. Thus, it is not very efficient with large datasets. In this study, we have used the multi-class support vector machines and systematic sampling with hierarchical clustering (SSHC-MCSVM) algorithm for gene expression data classification. The gene expression profiles are considered as large datasets. The gene expression datasets that are used in this study are two datasets for obese and lean individuals. In this proposed (SSHC-MCSVM) algorithm, the gene expression data are regrouped to new sets of genes based on systematic sampling with hierarchical clustering (SSHC) algorithm. The SSHC algorithm repeated n times and the k-partitions with clusters that have high adjusted Rand index (ARI) are chosen. The multi-class support vector machines are applied to the best regrouped gene expression data to classify the significant genes. The performance measures are accuracy, recall, and precision. The proposed algorithm which is SSHC-MCSVM could classify the significant genes with high accuracy, recall, and precision.

Data mining is a process which discovers patterns and retrieval knowledge in large datasets. Many learning and data mining algorithms rely on distance metrics. Cluster analysis is one of learning algorithms which adopted to biological data, for example, Microarray expression data. In this study, we assessed the validity of five distance metrics (Euclidean, Manhattan, Minkowski, Cosine and Mahalanobis) with the partitioning around medoids (PAM) algorithm on microarray datasets. Microarray datasets were pre-processed prior to analysis, and the evaluation of the algorithm was undertaken using Krzanowski-Lai validity index. Our results showed When selected microarray datasets were clustered with partitioning around medoids based on Manhattan distance, Minkowski, Cosine and Euclidean distance for different k partitions all exhibited unsatisfactory performance, however, the partitioning around medoids algorithm generates an optimal cluster solution when used with Mahalanobis distance.