Metin Madenciliği ile Tıbbi Tedavi Alanlarının Yakınlıklarının Ölçülmesi

Abstract

The fact that some of the symptoms are related to many medical treatment areas causes patients to have difficulty in making an appointment for treatment. In this study, 13 medical fields and 204 symptoms, which are available on the website of many public hospitals associated with T.C. Ministry of Health and used to help patients choose the right medical treatment branch according to their symptoms, were examined using text mining and data science techniques. Based on the content of the text used, the closeness among the medical treatment areas was calculated and the words and symptoms confusing the patients the most while deciding the treatment area were determined. When analyzing the words, meaningless words were ignored, and a word cloud was created on the symptoms of the diseases. In order to calculate the closeness of medical fields, 13x186 binary data was created, indicating whether each symptom exists. Later, the medical fields on this data set were clustered according to the symptoms using agglomerative hierarchical clustering algorithms and the proximity of medical treatment fields was found. In the results, the words that challenge patients the most and the text-based affinities of medical fields are shared. Reorganizing content of the official document used on the hospital websites using the results obtained on this study will help to reduce the number of appointments received from the wrong medical branches.