An extraction of medical information based on human handwritings

Nowadays communities are enabled to write and share their health experiences by Internet media, and a considerable number of experts and general public alike utilize them in the knowledge of medical analysis. Therefore, there are increasing needs to process some human handwriting in medical environment. This research, we applied the algorithm used to make a documentation extraction of the human handwriting automatically. The documentation category of this paper made for the research is about diabetics disease. We conducted an extraction of documents by using a term frequency - inversed document frequency method and measured the levels of resemblance by using a vector space model. The result of percentage found from 56 human handwritings was 81.818% and the unsuitable was 18.181%.