Supervised Text Classification for Marketing Analytics

Descripción

Build reliable supervised classifiers from marketing text and defensible human labels. Learners create coding rules, reconcile coders into gold-standard labels, transform text into predictive features, train a regularized elastic-net model, separate training from validation evidence, and use errors and learning curves to judge model quality and the value of collecting more labeled data.