Research Article
An Interactive Analysis of the Costs of Technologically Assisted Medical Reviews
@INPROCEEDINGS{10.4108/eai.20-4-2018.2276346, author={Giorgio Maria Di Nunzio}, title={An Interactive Analysis of the Costs of Technologically Assisted Medical Reviews}, proceedings={12th EAI International Conference on Pervasive Computing Technologies for Healthcare -- Demos, Posters, Doctoral Colloquium}, publisher={EAI}, proceedings_a={PERVASIVEHEALTH - EAI}, year={2018}, month={8}, keywords={systematic reviews probabilistic retrieval models interactive machine learning}, doi={10.4108/eai.20-4-2018.2276346} }- Giorgio Maria Di Nunzio
Year: 2018
An Interactive Analysis of the Costs of Technologically Assisted Medical Reviews
PERVASIVEHEALTH - EAI
EAI
DOI: 10.4108/eai.20-4-2018.2276346
Abstract
The understanding of medical data by experts with highly specific skills, as for example clinicians (a general term that encompasses every medical position involving patients), is necessary for diagnosing illnesses and administering treatment to patients through either medicine or protocols. However, the huge quantity of data produced by digital health technologies make the life of clinicians harder in terms of keeping up with such amount of information.
In this demo, we present an interactive application which monitors the amount of relevant literature found by a clinician in a Continuous Active Learning (CAL) framework. The application allows to study the actual costs of completing a systematic review within a 95% confidence interval by alternating random samples of documents with examples selected by a probabilistic machine learning approach.