Text Mining

Using text mining strategies, we developed a method to automatically classify lay requests to an Internet medical expert forum about involuntary childlessness.

We used a large sample of requests in the section "Wish for a Child" on the German website Rund-ums-Baby, which provides information for parents and potential parents. We developed a scoring procedure that calculated for any medical request the probability to belong to a certain subject matter (e.g. hormones or insemination) and to belong to a respective expectation of the sender (e.g. general information or interpretation of laboratory data).

On basis of this procedure we now develop a method how to find the "nearest neighbors" for the requests studied. If this procedure leads to satisfying results, it will be possible to classify new incoming requests to the expert forum and to look for the nearest neighbors of this request in the large stock of the expert forum. These nearest neighbors and the experts' former answers to these requests could be the basis of an automatic answer.


Ansprechpartner / Kontakt
Wolfgang Himmel
Prof. Dr. Wolfgang Himmel
+49 (0) 551 39-22648
+49 (0) 551 39-9530
CV (Deutsch/English)