Table 2.
Archives and Channels Doors Definitions | |||
---|---|---|---|
Definition | Description | Type of Data Extraction | AI Technologies and Automatisms Performed |
Platform based on ontology | Platform in use in our hospital for standardized data collection (BLADE, RedCAP, etc.). In this platform it is integrated a shared ontology that codifies data in unique, non-ambiguous way. | Organized and “ontologized” data | NEURAL NETWORKS |
Datawarehouse used in Fondazione Policlinico Gemelli IRCCS | Data warehouses in use in our hospital for clinical assistance (SI, Aria, Speed (advanced evolution of Spider [10], Armonia, TrackCare, etc.). In these systems, data are codified based on clinical practice (e.g., Hb value, date of surgery, etc.), and are data validated by conventional clinical use | Organized, not “ontologized” data | NEURAL NETWORKS |
Text mining extraction from PDF documents or electronic reports | All the electronic documents present in previous archives in which a procedure of text-mining extraction was applied to recover non-structured data. This is a very relevant part of data extraction, because we can recover a big quantity of granular information and translate it into structured data for usage in clinical practice and research. |
Not organized, not “ontologized” data | TEXT MINING AUTOLEARN NEURAL NETWORKS |