After applying the clustering method to a collection of text data, we may, fortunately, end up with several clusters. However, if there is no theoretical categorization scheme available for the data, how should we interpret the clusters? Should we do human-encoding? If so, how to evaluate the human interpretation of the clusters?
After applying the clustering method to a collection of text data, we may, fortunately, end up with several clusters. However, if there is no theoretical categorization scheme available for the data, how should we interpret the clusters? Should we do human-encoding? If so, how to evaluate the human interpretation of the clusters?