Main concepts in machine learning

Research output: Chapter in Book/Report/Conference proceedingChapter

6 Citations (Scopus)

Abstract

The last decade has seen a surge in machine learning studies in psychiatric and neurological disorders. Given its translational potential, machine learning is also capturing the interest of clinicians and other mental health practitioners. It is therefore important that these research and clinical communities develop a good appreciation of the nature of the machine learning process. This chapter aims to provide an accessible introduction to the main steps of the supervised learning pipeline, the most prevalent type of machine learning in the neuroscientific literature. The main stages of the standard pipeline include problem formulation, data preparation, feature engineering, model training, model evaluation, and post hoc analysis. Building a successful model is often an iterative process of adjustment of these several components. A good understanding of the rationale and challenges of each one is essential to avoid spurious interpretations.

Original languageEnglish
Title of host publicationMachine Learning
Subtitle of host publicationMethods and Applications to Brain Disorders
PublisherElsevier
Pages21-44
Number of pages24
ISBN (Electronic)9780128157398
ISBN (Print)9780128157398
DOIs
Publication statusPublished - 1 Jan 2019

Keywords

  • Classification
  • Cross-validation
  • Feature engineering
  • Model evaluation
  • Model training
  • Neurological disorders
  • Pipeline
  • Psychiatric disorders
  • Regression
  • Supervised machine learning

Fingerprint

Dive into the research topics of 'Main concepts in machine learning'. Together they form a unique fingerprint.

Cite this