chessboard.api.postsum.generatePointEstimate
- chessboard.api.postsum.generatePointEstimate(data, minprob, force_k=None)
Generate a clustering point estimate from MCMC samples.
When a single tile configuration is of interest, the MCMC samples from CHESSBOARD need to be summarized into a point estimate. This method clusters the pairwise sample clustering probability matrix (i.e. the probability that any 2 samples are clustered together across MCMC samples) using heirarchical clustering where k is the median number clusters across sampled tile configurations to obtain sample assignments. LSV assignments are obtained by averaging the marginal probabilities of each matrix entry being in the signal group across all samples in a cluster for the given LSV. Posteriors of the distributional parameters are updated based on sample and LSV assignments.
- Parameters
data (
Data) – Chessboard object. Must contain CHESSBOARD output after usingchessboard.api.io.Data.addChessboardOutput().minprob – The minimum probability of the average margin to be considered signal.