This is the same as the last PR, but without the columns of Twitter data on the annotations.
simple R code to compute co-annotation counts and display like a confusion matrix and report marginal probabilities per category
annotation data in long format .csv; coding standard in .csv
Stan code for the Dawid-Skene model with mixed panel data (not every annotator needs to label every item exactly once) and fixed priors (because there are so few annotators)
storopoli
commented
3 years ago
This is the same as the last PR, but without the columns of Twitter data on the annotations.
simple R code to compute co-annotation counts and display like a confusion matrix and report marginal probabilities per category
annotation data in long format .csv; coding standard in .csv
Stan code for the Dawid-Skene model with mixed panel data (not every annotator needs to label every item exactly once) and fixed priors (because there are so few annotators)
R code to fit the Dawid-Skene model to the data