A Clinical Text Classification Paradigm Using Weak Supervision…

This work develops a rule-based NLP algorithm to automatically generate labels for the training data, and then use the pre-trained word embeddings as deep representation features for training machine learning models.