Using Fuzzy C-means Clustering Algorithm to Diagnose the Severity of Anxiety

Diagnosing anxiety in the early stages by psychiatrists is one of the important steps in preventing and controlling these types of disorders. This study endeavors to present a method to diagnose the severity of anxiety using fuzzy C-means clustering algorithm. Moreover, the influence of each feature on measuring anxiety and clustering of clients is determined.

This study is quantitative and descriptive. A dataset including 300 clients related to three psychiatric clinics in Tehran was provided based on the Beck Anxiety Inventory questionnaire. Then, fuzzy C-means clustering algorithm segmented clients and determined the severity of their anxiety in each cluster. Additionally, this algorithm was employed for each feature, separately.

The psychiatric clinics' clients were divided into four clusters with the labels including no anxiety, minimal, moderate, and severe anxiety. Fuzzy C-means clustering algorithm diagnosed the anxiety of clients with 90.66% accuracy. Moreover, as a result of implementing the algorithm on each feature, the influence of the features on measuring anxiety and clustering of clients was determined.

Fuzzy C-means clustering algorithm diagnosed the anxiety of clients with a high accuracy. Segmenting patients by clustering approach and based on the important features, can be a dependable instrument for psychiatrist to make a decision in diagnosing the severity of anxiety in the early stages.