E-MFDBSCAN: an evolutionary clustering algorithm for gene expression time series

Abstract

DNA microarray experiments are frequently used because they have various advantages. However, gene expression data from DNA microarray experiments are noisy, and, consequently, the computations that are based on such noisy data may lack accuracy. In this paper, an evolutionary uncertain data-clustering algorithm, E-MFDBSCAN, and a prediction model using E-MFDBSCAN for uncertain data are proposed. The proposed methodology may be successfully applied to noisy gene expression data. In this methodology, global patterns of time series data can be extracted using our evolutionary clustering approach. These patterns are used to infer future projections. In the proposed methodology, an autoregressive time series function (using these patterns) used to predict the similarities among sets of gene expression clusters is constructed. The algorithms are tested with two different gene expression time series datasets.

Tanimoto Coefficient Similarity based Mean Shift Gentle Adaptive Boosted Clustering for Genomic Predictive Pattern Analytics

“…Tight clustering algorithm was employed in [19] to minimize time complexity of large microarray gene expression data. An evolutionary uncertain data-clustering algorithm was designed in [20] to determine the similarities among sets of gene expression clusters.…”