Replication Data for: Bespoke NPO Taxonomies - Step 03: Classification with Bootstrapping

Inputs and outputs of application of Naive Bayes text classifier to nonprofit mission statements to determine model performance in predicting organizational classes. The .RData files are quanteda corpus objects that the quanteda.textmodels implementation of the Naive Bayes classifier is applied to. The .rds files are the output of the classification, consisting of fit and performance metrics for that prediction class and dataset across all training dataset sizes (4,000 to 80,000 observations, in increments of 4,000), iterated 100 times at each size. For additional information on classification steps, please go to the project's replication GitHub page.