Figure 1.

Flowchart from the acquisition of the five tables containing the electronic health records (dark gray) in the training set that were merged into a single table (dark blue); preprocessed (gray) with the specification of what was removed (pink) prior to the creation of the document-term matrix (DTM) (yellow); the computation of the weights (light blue); the dimensionality reduction, that is the reduction of the terms used (light gray), and the final DTM used (green). DTM: document-term matrix; SOAP: symptoms, objectivity, diagnosis, or prescriptions; TF-iDF: term frequencies–inverse document frequencies.