![]() Coordinated highlighting of different representations for the same (or similar) dataset helps users reveal intersection and difference of those representations. In existing information visualization tools, brushing and linking techniques ( Becker and Cleveland, 1987) were used to show concordance. Similarly, a web resource can be mapped to multiple categories in the Open Directory ( Treating each category as a set containing many elements, concordance analysis to see how individual data elements are distributed in the categories helps users identify important categories as well as unveil the features of unknown elements. For example, a gene or a protein can be related to many gene ontology categories. A detailed example will be presented in section 5 to show how set concordance analysis helps compare clustering results.Īnother example is when one data element can be classified into multiple categories. For example, if we look at all sets together from the two clustering algorithms, the concordance can be checked by how many sets from one clustering result are similar to those from the other result. ![]() ![]() Comparison between two clustering algorithms can be formulated as a set concordance problem since most clustering algorithms generate disjoint sets (or clusters). Therefore, using only one clustering algorithm could bias the result. Different algorithms might come up with very different patterns depending on how they detect clusters. Biologists may use clustering algorithms to identify important patterns in the acquired dataset. Similar problems occur after the data acquisition step. While this can help significantly lower false positive rates, there is no interactive visualization tool for this purpose yet. Using set operations and various concordance measures to compare result sets from different signal algorithms, biologists can identify concordant/discordant genes across different signal algorithms. Subsequent analyses such as identification of important genes and power analysis depend on these numerical values. probe set signal algorithms) to acquire numerical values from microarrays, which represent gene activities ( Seo et al., 2006). For example, for an Affymetrix microarray experiment, biologists can use several algorithms (e.g. Since the choice of a data acquisition method can profoundly change the result values, it is crucial to check the concordance of the results from different acquisition methods to avoid high false positive rates. When scientists run an experiment, there may be several semi-standard methods (or algorithms) to acquire numerical values from a measurement device. This paper presents an interactive visualization tool called ConSet that enables researchers to visually analyze concordance of different empirical analyses that produce sets. Researchers can have a more judicious view on their research problems by comparing different analysis results on the same data set. Visualization of the concordance or discordance of different empirical analysis methods can help integrate important knowledge from different perspectives. We observed that ConSet enabled users to complete more tasks with fewer errors than the traditional interface did and most users preferred ConSet. A qualitative user study was conducted to evaluate how our tool works in comparison with a traditional set visualization tool based on a Venn diagram. Not only do we use a standard Venn diagram, we also introduce a new diagram called Fan diagram that allows users to compare two or three sets without any inconsistencies that may exist in Venn diagrams. ![]() ConSet provides an overview using an improved permutation matrix to enable users to easily identify relationships among sets with a large number of elements. In this paper, we present an interactive visualization tool called ConSet, where users can effectively examine relationships among multiple sets at once. To reach a more judicious conclusion, it is crucial to consider different perspectives by checking concordance among those result sets by different methods. For example, different scientific analysis methods with the same samples often lead to different or even conflicting conclusions. Scientific problem solving often involves concordance (or discordance) analysis among the result sets from different approaches.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |