Doctors diagnose patients for depression based on the way someone moves, sleeps, interacts, speaks, and their expressions. If depression goes undiagnosed, people will go without treatment. What if technology was able to see the signs first based on what is written on social media? A Clinical Psychological Science study was used to accurately predict depression based on what someone has written online.
A computerized text analysis allows being able to process large data in only minutes. This text analysis can spot the linguistic features that humans may miss. The percentage frequency of words and classes of words, lexical diversity, average sentence length, grammatical patterns, and others can help. By using personal essays, diary entries, and spoken word by people who have been known to be depressed like Kurt Cobain or Sylvia Plath will help detect a pattern and the clear differences in how people who are and are not depressed use speech.
Content relates to the meaning or subject matter statements. People who are depressed use negative emotion words like “lonely,” “sad,” or “miserable.” Those with depression also use more first person singular pronouns like “me,” “myself,” and “I” and fewer second and third person words. These results show how people who are depressed focus more on themselves and less with other people. The style of writing relates to expressing ourselves. Based on 64 different online mental health forums, absolutist words are used that mean absolute magnitude or probabilities like “always,” “nothing,” or “completely.” Their worldview is very black and white. In nineteen different control forums, absolutist words are in 50% of anxiety and depression forums and in 80% of suicide forums. Negative emotion words are less used in suicidal forums compared to anxiety and depression forums.
Because depressive episodes can happen again, absolutist words may play a role in depression episodes without having any symptoms. By combining automated text analysis with machine learning, we would be able to classify certain conditions from natural language text samples like blogs. As more data is provided and sophisticated algorithms are developed, machine learning will improve as well. Computers will help find specific subcategories for mental health like self-esteem, perfectionism, and social anxiety. Depression has increased up to 18% ever since 2005. More tools available to detect the early signs of depression will mean that a person’s mental health can be improved and will prevent the suicide rates of going up as well.
Located in downtown Midland, The Springboard Center’s mission is to offer programs and services to treat alcohol and drug addiction treatment using an evidence based curriculum, 12 step programs, diet, nutrition, exercise, emotional, mental and spiritual development for a long recovery. For more information, please call us at 432-620-0255 as we are open 24 hours a day, 7 days a week.