Growth Models for Categorical Response Variables: Standard, Latent-Class, and Hybrid Approaches

2017 ◽  
pp. 139-158
Author(s):  
Jeroen K. Vermunt
Keyword(s):  
Latent Class ◽  
Growth Models ◽  
Hybrid Approaches ◽  
Categorical Response ◽  
Response Variables

scholarly journals Towards the characterization and validation of alcohol use disorder subtypes: integrating consumption and symptom data

2013 ◽  
Vol 44 (1) ◽  
pp. 143-159 ◽  
Author(s):  
K. M. Jackson ◽  
K. K. Bucholz ◽  
P. K. Wood ◽  
D. Steinley ◽  
J. D. Grant ◽  
...  
Keyword(s):  
Alcohol Consumption ◽  
Alcohol Use ◽  
Alcohol Use Disorder ◽  
Model Comparison ◽  
Latent Class ◽  
Common Factor ◽  
Hybrid Approaches ◽  
Epidemiologic Survey ◽  
Alcohol Involvement ◽  
Best Fitting

BackgroundThere is evidence that measures of alcohol consumption, dependence and abuse are valid indicators of qualitatively different subtypes of alcohol involvement yet also fall along a continuum. The present study attempts to resolve the extent to which variations in alcohol involvement reflect a difference in kindversusa difference in degree.MethodData were taken from the 2001–2002 National Epidemiologic Survey of Alcohol and Related Conditions. The sample (51% male; 72% white/non-Hispanic) included respondents reporting past 12-month drinking at both waves (wave 1:n = 33644; wave 2:n = 25186). We compared factor mixture models (FMMs), a hybrid of common factor analysis (FA) and latent class analysis (LCA), against FA and LCA models using past 12-month alcohol use disorder (AUD) criteria and five indicators of alcohol consumption reflecting frequency and heaviness of drinking.ResultsModel comparison revealed that the best-fitting model at wave 1 was a one-factor four-class FMM, with classes primarily varying across dependence and consumption indices. The model was replicated using wave 2 data, and validated against AUD and dependence diagnoses. Class stability from waves 1 to 2 was moderate, with greatest agreement for the infrequent drinking class. Within-class associations in the underlying latent factor also revealed modest agreement over time.ConclusionsThere is evidence that alcohol involvement can be considered both categorical and continuous, with responses reduced to four patterns that quantitatively vary along a single dimension. Nosologists may consider hybrid approaches involving groups that vary in pattern of consumption and dependence symptomatology as well as variation of severity within group.

Zero-Inflated Latent Class Growth Models: A Tutorial Using Data From a Longitudinal Cohort

2013 ◽  
Author(s):  
Trynke Hoekstra ◽  
Sterling M. Mcpherson ◽  
Celestina Barbosa-Leiker ◽  
Jos W. R. Twisk
Keyword(s):  
Latent Class ◽  
Growth Models ◽  
Longitudinal Cohort ◽  
Using Data ◽  
Latent Class Growth

scholarly journals Trajectories of Depressive Symptoms Experienced by Older People: Effects of Time, Hurricane Sandy, and the Great Recession

2020 ◽  
Author(s):  
Rachel Pruchno ◽  
Maureen Wilson-Genderson ◽  
Allison Heid ◽  
Francine Cartwright
Keyword(s):  
Depressive Symptoms ◽  
Great Recession ◽  
Traumatic Stress ◽  
Latent Class ◽  
Growth Models ◽  
Multinomial Logistic Regression ◽  
Hurricane Sandy ◽  
The Great Recession ◽  
Symptom Trajectories ◽  
History Of

Abstract Objectives To examine depressive symptom trajectories as a function of time and exposure to Hurricane Sandy, accounting for the effects of the Great Recession. Methods We analyzed 6 waves of data from a 12-year panel using latent class growth models and multinomial logistic regression. Results We identified 4 groups of people experiencing different trajectories of depressive symptoms. The groups differed on baseline characteristics (gender, age, education, income, race), history of diagnosed depression, and initial level of depressive symptoms. The group with the highest levels of depressive symptoms reported greater levels of peri-traumatic stress exposure to Hurricane Sandy. Discussion Depressive symptoms increased as a function of the Great Recession, but exposure to Hurricane Sandy was not associated with subsequent increases in depressive symptoms for any of the 4 groups. People who consistently experienced high levels of depressive symptoms over time reported the highest levels of peri-traumatic stress during Hurricane Sandy. Findings highlight the importance of accounting for historical trends when studying the effects of disaster, identify people likely to be at risk during a disaster, and provide novel information about the causal relationship between exposure to disaster and depressive symptoms.

scholarly journals Support Vector Machines and Support Vector Regression

2022 ◽  
pp. 337-378
Author(s):  
Osval Antonio Montesinos López ◽  
Abelardo Montesinos López ◽  
Jose Crossa
Keyword(s):  
Support Vector Machines ◽  
Plant Breeding ◽  
Genomic Prediction ◽  
Kernel Functions ◽  
Support Vector ◽  
Continuous Response ◽  
Support Vector Classifier ◽  
Vector Machines ◽  
Categorical Response ◽  
Response Variables

AbstractIn this chapter, the support vector machines (svm) methods are studied. We first point out the origin and popularity of these methods and then we define the hyperplane concept which is the key for building these methods. We derive methods related to svm: the maximum margin classifier and the support vector classifier. We describe the derivation of the svm along with some kernel functions that are fundamental for building the different kernels methods that are allowed in svm. We explain how the svm for binary response variables can be expanded for categorical response variables and give examples of svm for binary and categorical response variables with plant breeding data for genomic selection. Finally, general issues for adopting the svm methodology for continuous response variables are provided, and some examples of svm for continuous response variables for genomic prediction are described.

Sign in / Sign up

Export Citation Format

Share Document