Control GAD (Generalized Anxiety Disorder)

• Mercy Falconi

• Huridocs

Generalized anxiety disorder (GAD) is one of these disorders, characterized by the presence of excessive worries and alertness over different issues; people with GAD are constantly anticipating some kind of disasters about health, family, work, studies, etc. This model describes if it is possible to identify the presence of anxiety in the written expressions of a person, and with this information determining the frequency of worrying thoughts, in order to diagnose the existence of GAD. The results of our models confirm the hypothesis about the capacity of an AI model to predict the level of anxiety in a textual phrase and supports the virtual care of people with this condition. The analysis consists in to classify the text in a set of levels of anxiety, by means of a deep learning model, trained with phrases tagged with an anxiety level, according to the Patient Health Questionnaire (PHQ-8) score of a group of persons in a clinical experiment. Word embeddings technique over windowed sequences was applied to train the model, with vector representations for the meaning of each word and then the variants of recurrent neural network with word embedding were used for prediction.

Input variables : Textual Phrases from Interview Transcripts
Output Variables : Anxiety level present in the text -none, mild, moderate, moderately severe, severe

Visit Model : github.com

Additional links : medium.com