statGraph is an R package that compiles several statistical methods for graphs (networks). For example, there are methods for graph parameter estimation, model selection (GIC - Graph Information Criterion), a statistical test to discriminate two or more populations of graphs (GANOVA - Graph Analysis of Variability), correlation between graphs, etc.
Package source: statGraph_0.01-1.tar.gz
Windows binary: statGraph_0.01-1.zip
Mac OS X binary: statGraph_0.01-1.tgz
- Takahashi DY, Sato JR, Ferreira CE, Fujita A. Discriminating different classes of biological networks by analyzing the graphs spectra distribution. PLoS ONE. 7: e49949, 2012.
- de Siqueira Santos S, Takahashi DY, Sato JR, Ferreira CE, Fujita A. Statistical methods in graphs: parameter estimation, model selection, and test. In: Yongtang Shi, Sueliang Li, Matthias Dehmer (Eds.). Mathematical foundations and applications of graph entropy. Wiley-Blackwell (in press)
Slope is a non-parametric and data-driven method for estimating the number of clusters. This approach uses the output of any clustering algorithm.
Package source: slope_0.01-1.tar.gz
Windows binary: slope_0.01-1.zip
Mac OS X binary: slope_0.01-1.tgz
- Fujita A, Takahashi DY, Patriota AG. A non-parametric method to estimate the number of clusters. Computational Statistics and Data Analysis. 73: 27 - 39, 2014.
ANOCVA statistically tests whether two or more populations are equally clustered. ANOCVA allows the comparison of multiple groups simultaneously and identifies features that contribute to the differential clustering.
Package source: anocva_0.1.0_R_x86_64-pc-linux-gnu.tar.gz
Windows binary: anocva_0.1.0.zip
Mac OS X binary: anocva_0.1.0.tar.gz
- Fujita A, Takahashi DY, Patriota AG, Sato JR. A non-parametric statistical test to compare clusters with applications in functional magnetic resonance imaging data. Statistics in Medicine. 33: 4949 - 4962, 2014.