Evaluate the loss of a point estimate of the partition compared to a gold standard according to a given loss function
evalClustLoss(c, gs, lossFn = "F-measure", a = 1, b = 1)the cost of the point estimate c in regard of the
gold standard gs for a given loss function.
vector of length n containing the estimated partition
of the n observations.
vector of length n containing the gold standard
partition of the n observations.
character string specifying the loss function to be used. Either "F-measure" or "Binder" (see Details). Default is "F-measure".
only relevant if lossFn is "Binder". Penalty for wrong
co-clustering in c compared to gs. Defaults is 1.
only relevant if lossFn is "Binder". Penalty for missed
co-clustering in c compared to gs. Defaults is 1.
Boris Hejblum
The cost of a point estimate partition is calculated using either a pairwise coincidence loss function (Binder), or 1-Fmeasure (F-measure).
J.W. Lau & P.J. Green. Bayesian Model-Based Clustering Procedures, Journal of Computational and Graphical Statistics, 16(3): 526-558, 2007.
D. B. Dahl. Model-Based Clustering for Expression Data via a Dirichlet Process Mixture Model, in Bayesian Inference for Gene Expression and Proteomics, K.-A. Do, P. Muller, M. Vannucci (Eds.), Cambridge University Press, 2006.
similarityMat, cluster_est_binder