I have created a model for text analysis using tokens , ngram and used it for classification on text . in each run the model give different results (mostly wrong values) Why is this happening ?
can you share a reproducible example?
Maybe the model that is used has a random initialization step.
Could you try prepending the model run with a set.seed() call, e.g. set.seed(123)?
set.seed()
set.seed(123)
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.