Classification/Segmentation
Some describe segmentation as simply classifying an entire population based on a few levels of a single factor; I would see it more from a machine learning perspective - it can be as simple as using a logistic regression to differentiate a population based on covariates, or as complex as clustering models based on distance matrices or latent measurement models. The statistical segmentation techniques include support vector machines (SVMs), K-means, classification trees, latent class analysis and more.
A Latent Class Analysis to Support New Diagnose Criteria for Gambling Disorder in DSM5 - A Bayesian Approach
The plot resulted from a LCA shows that the "Illegal Acts" criterion has a low probability of being met among patients with gambling disorder, which in another way supports the validity of eliminating this criterion in the new DSM5. The ultimate result for LCA is to give the probability for a person to be classified in a certain class, based on Bayes rules.
The ROC Curve and the Corresponding AUC of Two SVM Models with Different Number of Predicting Features
3D Representation of the Space of Common Psychiatric Disorders in the NCS-A
1. Major Depressive Episode/Dysthymia; 2. Generalized Anxiety Disorder; 3. Mania/Hypomania; 4. Specific Phobia; 5. Agoraphobia (with/without panic disorder); 6. Social Anxiety Disorder; 7. Panic Disorder; 8. Separation Disorder; 9. Post-traumatic Stress Disorder; 10. Eating Disorder; 11. Attention Deficit Hyperactivity Disorder; 12. Oppositional Defiant Disorder; 13. Conduct Disorder; 14. Drug Use Disorder; 15. Alcohol Use Disorder; 16. Nicotine Dependence.
Classification based on Bayesian Networks