Abstract-In this article, we evaluate the performance of three clustering algorithms, hard K-Means, single linkage, and a simulated annealing (SA) based technique, in conjunction with four cluster validity indices, namely Davies-Bouldin index, Dunn's index, Calinski-Harabasz index, and a recently developed index I . Based on a relation between the index I and the Dunn's index, a lower bound of the value of the former is theoretically estimated in order to get unique hard K-partition when the data set has distinct substructures. The effectiveness of the different validity indices and clustering methods in automatically evolving the appropriate number of clusters is demonstrated experimentally for both artificial and real-life data sets with the number of clusters varying from two to ten. Once the appropriate number of clusters is determined, the SA-based clustering technique is used for proper partitioning of the data into the said number of clusters.
This article describes a simulated annealing based multi-objective optimization algorithm that incorporates the concept of archive in order to provide a set of trade-off solutions of the problem under consideration. To determine the acceptance probability of a new solution visa -vis the current solution, an elaborate procedure is followed that takes into account the domination status of the new solution with the current solution, as well as those in the archive. A measure of the amount of domination between two solutions is also used for this purpose. A complexity analysis of the proposed algorithm is provided. An extensive comparative study of the proposed algorithm with two other existing and well-known multi-objective evolutionary algorithms (MOEAs) demonstrate the effectiveness of the former with respect to five existing performance measures, and several test problems of varying degrees of difficulties. In particular, the proposed algorithm is found to be significantly superior for many-objective test problems (e.g., 4, 5, 10 and 15 objective problems), while recent studies have indicated that the Pareto ranking-based MOEAs perform poorly for such problems. In a part of the investigation, comparison of the real-coded version of the proposed algorithm is conducted with a very recent multiobjective simulated annealing algorithm where the performance of the former is found to be generally superior to that of the latter.
BackgroundMicroRNAs are a class of small noncoding RNAs that are abnormally expressed in different cancer cells. Molecular signature of miRNAs in different malignancies suggests that these are not only actively involved in the pathogenesis of human cancer but also have a significant role in patients survival. The differential expression patterns of specific miRNAs in a specific cancer tissue type have been reported in hundreds of research articles. However limited attempt has been made to collate this multitude of information and obtain a global perspective of miRNA dysregulation in multiple cancer types.ResultsIn this article a cancer-miRNA network is developed by mining the literature of experimentally verified cancer-miRNA relationships. This network throws up several new and interesting biological insights which were not evident in individual experiments, but become evident when studied in the global perspective. From the network a number of cancer-miRNA modules have been identified based on a computational approach to mine associations between cancer types and miRNAs. The modules that are generated based on these association are found to have a number of common predicted target onco/tumor suppressor genes. This suggests a combinatorial effect of the module associated miRNAs on target gene regulation in selective cancer tissues or cell lines. Moreover, neighboring miRNAs (group of miRNAs that are located within 50 kb of genomic location) of these modules show similar dysregulation patterns suggesting common regulatory pathway. Besides this, neighboring miRNAs may also show a similar dysregulation patterns (differentially coexpressed) in the cancer tissues. In this study, we found that in 67% of the cancer types have at least two neighboring miRNAs showing downregulation which is statistically significant (P < 10-7, Randomization test). A similar result is obtained for the neighboring miRNAs showing upregulation in specific cancer type. These results elucidate the fact that the neighboring miRNAs might be differentially coexpressed in cancer tissues as that of the normal tissue types. Additionally, cancer-miRNA network efficiently detect hub miRNAs dysregulated in many cancer types and identify cancer specific miRNAs. Depending on the expression patterns, it is possible to identify those hubs that have strong oncogenic or tumor suppressor characteristics.ConclusionsLimited work has been done towards revealing the fact that a number of miRNAs can control commonly altered regulatory pathways. However, this becomes immediately evident by accompanying the analysis of cancer-miRNA relationships in the proposed network model. These raise many unaddressed issues in miRNA research that have never been reported previously. These observations are expected to have an intense implication in cancer and may be useful for further research.
An important approach for unsupervised landcover classification in remote sensing images is the clustering of pixels in the spectral domain into several fuzzy partitions. In this paper, a multiobjective optimization algorithm is utilized to tackle the problem of fuzzy partitioning where a number of fuzzy cluster validity indexes are simultaneously optimized. The resultant set of near-Pareto-optimal solutions contains a number of nondominated solutions, which the user can judge relatively and pick up the most promising one according to the problem requirements.
Real-coded encoding of the cluster centers is used for this purpose.Results demonstrating the effectiveness of the proposed technique are provided for numeric remote sensing data described in terms of feature vectors. Different landcover regions in remote sensing imagery have also been classified using the proposed technique to establish its efficiency.Index Terms-Cluster validity measures, fuzzy clustering, genetic algorithm (GA), multiobjective optimization (MOO), Pareto-optimal, pixel classification, remote sensing imagery.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.