Open ococrook opened 4 years ago
tagmPredict gives and error when using NoveltyTAGM results instead of usually MCMC results:
tanTAGMnov <- tagmMcmcTrain_Nov(object = tan2009r1, fcol = "markers", numIter = 100, # typically needs to be 15,000+ thin = 2, #typically about 10 burnin = 20, # typically about 10,000 numChains = 1, # chais in parraellel recommend 4+ beta0 = 1, # recommend 1 or 1/number of proteins K_bar = 10) # number of new classes allowed tanTagmNoveltyRes <- tagmNoveltyProcess(object = tan2009r1, params = tanTAGMnov) tanTagmNoveltyparams <- tagmMcmcProcess(tanTAGMnov) tan2009r1 <- tagmPredict(object = tan2009r1, params = tanTagmNoveltyparams, probJoint = TRUE)
Error in dimnames(x) <- dn : length of 'dimnames' [2] not equal to array extent
The offending lines in tagmMcmcpreidct are
K <- chains(params)[[1]]@K .joint <- rbind(params@summary@tagm.joint, .probmat)
These need to replaced with
K <- length(markers) .joint <- rbind(params@summary@tagm.joint[, markers], .probmat)
tagmPredict gives and error when using NoveltyTAGM results instead of usually MCMC results:
Error in dimnames(x) <- dn : length of 'dimnames' [2] not equal to array extent
The offending lines in tagmMcmcpreidct are
These need to replaced with