Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Bruno Iochins Grisci, Marcio Dorn (Structural Bioinformatics and Computational Biology Lab - Institute of Informatics - Federal University of Rio Grande do Sul, Porto Alegre, Brazil)
This image shows several 2D projections of a high-dimensional data set of gene expression from healthy samples (blue dots) and liver cancer samples (pink dots). The scatter plot in the center shows the original data using t-SNE. Each smaller plot on the circle represents a gene selection using a specific feature selection algorithm and were created using "weighted t-SNE," a method proposed by us that inputs each dimension's relevance score into the distance metric. The circle is separating the centroids of each class for each small plot (the squares). Using weighted t-SNE, it is possible to visually inspect which algorithms are producing selections better able to divide the samples into the two classes. The data set can be downloaded from the CuMiDa database (https://sbcb.inf.ufrgs.br/cumida). The algorithms shown are Linear SVM, MRMR, Mutual Information, Random Forest, ReliefF, Genetic Algorithm and ReliefF, Genetic Algorithm and SVM, SVM-RFE, Decision Tree, Kruskal-Wallis filter, and Lasso.