Search

Search Results

• Classification of depression-related online forums using Natural Language Processing

  181-208.

  Mate Fanni

  Views:

  70

  The study of the phenomenon of depression is not new in sociology, but since the depression
  is becoming a wider social problem, it is still a relevant issue today. In addition to the biomedical and psychological aspects of depression, the sociological perspective is becoming more
  noteworthy in the discourse about the causes of depression. In the research of the discourse
  on depression, the online texts offer many new possibilities, as the forum’s anonymity and
  accessability make the online seeking for help popular. In this research, natural language
  processing (logistic regression) was applied to find patterns in the definition of depression
  in lay discourses. These methods make it possible to analyze a large amount of text - which
  would have been difficult to process with human resources. During the analysis, 67 857 posts of
  English-speaking online forums were categorized along the categories of the scientific discourse
  about depression. This study presents the first results, which shows logistic regression classifier
  performs like the annotators. . Although the research has analyzed English-speaking forums, my
  findings may be useful to anyone observing abstract sociological concepts in online texts written
  by users.