An Automatic Approach for Detecting Cognitive Distortion from Spontaneous Thinking

Cognitive distortions are spontaneous thought processes that humans develop. A few research investigations investigated machine learning methods and tools for detecting and categorizing in the environment of cognitive-behavioral therapy. This research aims to detect the cognitive distortions from the spontaneous thinking of patient interaction. This work implements three models using natural language processing, feature extraction methods and a machine learning technique. The best-proposed model CDTS is implemented using TF-IDF with a support vector machine classifier named for cognitive distortion detection. The experiments have been performed on the CBT dataset using all three models and found the best results in terms of f1-score is 94% to detect cognitive distortion using the proposed CDTS model. This model is helpful in providing better cognitive behavioural therapy by psychiatrists and clinical psychologists with the help of early detection of cognitive distortion.