A Robust ISOMAP Algorithm And Its Application

DNA microarrays are used to produce large sets of expression measurements from which specific biological information is sought. And efficient and reliable algorithms for data visualization are required when analyzing microarray data. ISOMAP is a new algorithm of dimensionality reduction which can be used to data visualization. But the emergence of "shortcut" edges can make the corresponding neighborhood graph represent the neighborhood structure of the data falsely due to the noise or the unsuitable neighborhood size, thus ISOMAP cannot be applied successfully. Based on the fact that linked distance is sensitive to "shortcut" edges, this paper presents a new variant of ISOMAP, i.e. LD-ISOMAP(Linked distance ISOMAP),which can prune effectively "shortcut" edges existed possibly in the neighborhood graph and thus is much less sensitive to the neighborhood size and more topologically stable than ISOMAP. The effectively of LD-ISOMAP is verified by the experimental results very well.