Machine learning in obsessive-compulsive disorder medications

Abstract

Obsessive-compulsive disorder (OCD) is the fourth most common psychiatric disorder with a significant morbidity rate. Despite various treatment modalities and medications, some patients show no definitive response. The aim of this study is to classify the medications of OCD with machine learning (ML) methods and to compare the classification performances of the decision tree (DT), chi-square automatic interaction detection (CHAID) algorithm, and linear model in ML methods. This research is a descriptive analytical study based on co-word and artificial intelligence methods. The DT models were created with a target (total weight link strength). For hyperparameter optimization, the Gini index was used as the weight total link strength. The performance of the DT model was evaluated based on the prediction model. A total of 116 drugs were extracted from 6574 articles based on co-word analysis, and 56 drugs were classified as the DT's root. These drugs were categorized into six groups in the EWKM diagram. The DT was constructed using the weight.total.link index, with 7 items in Label 3 and 42 items in Label 5 serving as DT leaves. The ML analysis provided valuable insights into the efficacy of various medications such as clomipramine, duloxetine, and pindolol, as well as supplements such as folate, in the treatment of OCD. Treating concomitant diseases, namely hypothyroidism and streptococcal infection could improve the efficacy of treatment.

Keywords

• Machine learning
• Decision tree
• Obsessive-compulsive disorder
• OCD
• Clomipramine
• Duloxetine