Identifying cognitive distortion by convolutional neural network based text classification

• Main Authors: Xing, Zhenchang, Zhao, Xuejiao, Miao, Chunyan
• Other Authors: School of Computer Science and Engineering
• Format: Article
• Language: English
• Published: 2018
• Subjects: Cognitive Distortion; Word Embedding; DRNTU::Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/89482; http://hdl.handle.net/10220/47264
• Institution: Nanyang Technological University

Cognitive distortions have a way of playing havoc with our lives. The most important step to untwist the irrational thinking is identifying the forms of the cognitive distortion. The daily narration or diaries of the patients are always used by the cognitive-behavioral therapists as a clue to identify the cognitive distortion. But these natural language materials are always diverse and desultory which affect the efficiency and accuracy of identification. In this research, we propose a model called ICODLE (Identifying Cognitive Distortion by Deep Learning) which utilizes the daily narration or diaries of the patients to identify the forms of the cognitive distortion. ICODLE collect the daily narration and diaries from the authoritative books and webpages in CBT (Cognitive-Behavioral Therapy) domain. Then ICODLE creates the database of the 10 forms of cognitive distortion which were defined by David D. Burns. By utilizing the advanced deep learning techniques (e.g., Word Embedding, CNN (Convolutional Neural Network), etc.), ICODLE can identify the forms of the patients' cognitive distortions without the features extraction. ICODLE can effectively assist the patients and the cognitive-behavioral therapists to diagnose the cognitive distortions. ICODLE also benefit to build up the online persuasion system.