A valuable application of Term Selection is to create a model from a training set, with a known response, to utilize prediction formulas on a new set of data, without the response, for making predictions. The formula includes terms in a format showing an underscore with the term name. When creating a DTM for the new data, the format of the column names for terms in the term followed by the weighting scheme. One must update one or both of the formula references and term variables to allow for the predictions to work. It would be great to have a way to create model formulas and DTM in new data that immediately match in format. See the images below of the mismatch for clarity.
Another issue is when the DTM for the new data does not include all terms used by the training set prediction formula, and will not provide predictions due to missing terms. This can be corrected manually as well; however, it is time consuming.
Thanks for considering to update features to make term selection even better!
