6 Applying a Model to categorize Documents
Here we show how to generate and save a hold-out set, a wordlist and a model to use them for scoring on unseen documents. We do the text classification using a Support Vector Machine. Some of the additional techniques we show in this comprehensive video are weights for handling the imbalanced classes and how to combine a preprocessing model and a learner into on forecasting model. A key difference between usual classification and classification in text mining is that the wordlist from the model training needs to be added during the document preprocessing.