Text Mining for Personalized Knowledge Extraction From Online Support Groups

The traditional approach to health care is being revolutionized by the rapid adoption of patient‐centered healthcare models. The successful transformation of patients from passive recipients to active participants is largely attributed to increased access to healthcare information. Online support groups present a platform to seek and exchange information in an inclusive environment. As the volume of text on online support groups continues to grow exponentially, it is imperative to improve the quality of retrieved information in terms of relevance, reliability, and usefulness. We present a text‐mining approach that generates a knowledge extraction layer to address this void in personalized information retrieval from online support groups. The knowledge extraction layer encapsulates an ensemble of text‐mining techniques with a domain ontology to interpose an investigable and extensible structure on hitherto unstructured text. This structure is not limited to personalized information retrieval for patients, as it also imparts aggregates for crowdsourcing analytics by healthcare researchers. The proposed approach was successfully trialed on an active online support group consisting of 800,000 posts by 72,066 participants. Demonstrations for both patient and researcher use cases accentuate the value of the proposed approach to unlock a broad spectrum of personalized and aggregate knowledge concealed within crowdsourced content.