Natural language analysis diagnoses mental state with an accuracy rate of 83%

Big data is expected to strengthen the prediction of mental illness and improve the treatment and management of mental illness. According to reports, after analyzing the interviews of the risk groups of adolescents, the probability of mental illness can be predicted within 2 years with an accuracy of 83%. The other experiment is conducted for the general risk group with a prediction accuracy of 79%. Thus, speech analysis techniques can help predict psychological disorders such as mental illness.

One of the symptoms of mental illness is the lack of organization, which can be assessed through interviews. In the speech of a mentally ill person, there may be a lack of relevance or complexity of the topic, the content is not high, and if it is serious, it may be detrimental to communication. Before the outbreak of mental illness among young and dangerous groups, there is a long-term mild language barrier.

The Icahn School of Medicine (ISMMS) in Mount Sinai, USA, uses computerized automated interviews to determine the differences between people with signs of mental illness and normal people through automated natural language processing. Cheryl Corcoran, assistant professor of psychiatry, pointed out that the results are very exciting. The technology helps predict mental illness and can lock down cognitive deficits that lead to language barriers and prevent mental illness.

Broadly speaking, language and behavior are the main sources of information for psychiatrists to diagnose and treat mental illness. Although there are new computer algorithms to identify complex behaviors, words are easier to obtain and analyze. Technology can be widely used in psychiatry and even other medical fields.