Year: 2018
Type: conference-paper
Abstract: Health information technologies have greatly facilitated sharing of personal health data for secondary use, which is critical to medical and health research. However, there is a growing concern about privacy due to data sharing and publishing. Medical and health data typically contain unstructured text documents, such as clinical narratives, pathology reports, and discharge summaries. This study concerns privacy-preserving extraction, summary, and release of information from medical documents. Existing studies on privacy-preserving data mining and publishing focus mostly on structured data. We propose a novel approach to enable privacy-preserving extract, summarize, query and report patients’ demographic, health and medical information from medical documents. The extracted data is represented in a semi-structured, set-valued data format, which can be stored in a health information system for query and analysis. The privacy preserving mechanism is based on the cutting-edge idea of differential privacy, which offers rigorous privacy guarantee.
Authors: Xiaobai Li, Jialun Qin
Institution: University of Massachusetts Lowell
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W2787795726