JMIR Publications

ABSTRACT

Background:

ML is part of artificial intelligence (AI) that uses statistical techniques to give computer systems the ability to progressively build algorithms with data. ML, with its processing speed, higher volumes of data, and multitude number of sources, is challenging evidence based medicine. NLP, a domain of linguistic and AI, is essential in psychiatry since based deficits are commonly associated with mental illness. When combined ML with NLP creates automated systems which are emerging as valuable tools in psychiatric research.

Objective:

To identify clinical knowledge and perspectives in psychiatry regarding the combined use of machine learning (ML) and natural language processing (NLP) techniques.

Methods:

This systematic review follows PRISMA guidelines. Researches were performed in Pubmed, Scopus, Science direct, Psycinfo, Cocheranelibrary. This review investigated studies about the clinical use of ML with NLP, written in English, after 2010.

Results:

The selected studies were heterogeneous in terms of topic, objectives, and methods. ML-NLP increases the speed and specificity of diagnosis by extracting and classifying criteria from data which can help detect people at clinical risk for depression or suicide attempts. Additionally, it improves knowledge in psychophathology by establishing risk factors using statistical analysis; that leads to highly accurate prognosis which can aid in clinical decision making. Therefore, in accordance with the ethical principal of beneficience, guidelines should highlight the clinical benefits of ML-NLP techniques.

Conclusions:

ML in combination with NLP, establish valuable clinical patterns using high quantities of data, creating new paradigms in psychiatric research. Clinical Trial: Registered on PROSPERO (number CRD42019107376)