I've tried to run your toy example (codes as follows) with cellassign (0.99.21) without any error. However, I cannot get stable results under the default setting as shown in the code. In other words, the accuracy is 1 for some runs but less than 1 for some other runs. By reading other issues, I think this is because the algorithm ends with a local optimum. Could you please tell me what parameter I should change to stabilize the results? Thank you!
' @examples
' data(example_sce)
' data(example_marker_mat)
'
' fit <- em_result <- cellassign(example_sce[rownames(example_marker_mat),],
' marker_gene_info = example_marker_mat,
' s = colSums(SummarizedExperiment::assay(example_sce, "counts")),
To whom it may concern,
I've tried to run your toy example (codes as follows) with cellassign (0.99.21) without any error. However, I cannot get stable results under the default setting as shown in the code. In other words, the accuracy is 1 for some runs but less than 1 for some other runs. By reading other issues, I think this is because the algorithm ends with a local optimum. Could you please tell me what parameter I should change to stabilize the results? Thank you!
' @examples
' data(example_sce)
' data(example_marker_mat)
'
' fit <- em_result <- cellassign(example_sce[rownames(example_marker_mat),],
' marker_gene_info = example_marker_mat,
' s = colSums(SummarizedExperiment::assay(example_sce, "counts")),
' learning_rate = 1e-2,
' shrinkage = TRUE,
' verbose = FALSE)