Latent and observable variables: Difference between revisions

Revision as of 02:19, 20 December 2021

Latest revision as of 16:11, 19 May 2025

(39 intermediate revisions by 29 users not shown)

Line 1:

{{Short description|Variable not directly observed}}

In [[statistics]], '''latent variables''' (from [[Latin]]: [[present participle]] of ''lateo'' (“lie hidden”), as opposed to [[observable variable]]s) are [[Variable (mathematics)|variables]] that are not directly observed but are rather [[Statistical inference|inferred]] (through a [[mathematical model]]) from other variables that are observed (directly measured). Mathematical models that aim to explain observed variables in terms of latent variables are called [[latent variable model]]s. Latent variable models are used in many disciplines, including [[psychology]], [[demography]], [[economics]], [[engineering]], [[medicine]], [[physics]], [[machine learning]]/[[artificial intelligence]], [[bioinformatics]], [[chemometrics]], [[natural language processing]], [[econometrics]], [[management]] and the [[social sciences]].

{{For|similar uses|Hidden variable (disambiguation){{!}}Hidden variable}}

In [[statistics]], '''latent variables''' (from [[Latin]]: [[present participle]] of ''{{wikt-lang|la|lateo''}} (“lie{{gloss|lie hidden”),hidden}}{{cn|date=April as opposed to [[observable variable]]s2025}}) are [[Variable (mathematics)|variables]] that arecan not directly observed but areonly ratherbe [[Statistical inference|inferred]] (indirectly through a [[mathematical model]]) from other '''observable variables''' that arecan observedbe (directly [[observation|observed]] or [[measurement|measured)]].<ref>Dodge, MathematicalY. models(2003) that aim''The toOxford explain observed variables in termsDictionary of latentStatistical variablesTerms'', areOUP. called{{isbn|0-19-920613-9}}</ref> Such ''[[latent variable model]]s. Latent variable models'' are used in many disciplines, including [[psychologyengineering]], [[demographymedicine]], [[economicsecology]], [[engineeringphysics]], [[medicinemachine learning]], /[[physicsartificial intelligence]], [[machinenatural learninglanguage processing]]/, [[artificial intelligencebioinformatics]], [[bioinformaticschemometrics]], [[chemometricsdemography]], [[naturaleconomics]], language processing[[management]], [[econometricspolitical science]], [[managementpsychology]] and the [[social sciences]].

Latent variables may correspond to aspects of physical reality. These could in principle be measured, but may not be for practical reasons. In this situation, the term ''hidden variables'' is commonly used (reflecting the fact that the variables are meaningful, but not observable). Other latent variables correspond to abstract concepts, like categories, behavioral or mental states, or data structures. The terms ''hypothetical variables'' or ''hypothetical constructs'' may be used in these situations.

Latent variables may correspond to aspects of physical reality. These could in principle be measured, but may not be for practical reasons. Among the earliest expressions of this idea is [[Francis Bacon]]'s [[polemic]] the ''[[Novum Organum]]'', itself a challenge to the more traditional logic expressed in [[Aristotle]]'s [[Organon]]:

{{quote|But the latent process of which we speak, is far from being obvious to men’s minds, beset as they now are. For we mean not the measures, symptoms, or degrees of any process which can be exhibited in the bodies themselves, but simply a continued process, which, for the most part, escapes the observation of the senses.|[[Francis Bacon]], ''[[Novum Organum]]''<ref>{{cite book|author-last=Bacon|author-first=Francis|title=Novum Organum|chapter=APHORISMS—BOOK II: ON THE INTERPRETATION OF NATURE, OR THE REIGN OF MAN|url=https://www.gutenberg.org/files/45988/45988-h/45988-h.htm}}</ref>

}}

Latent variables may correspond to aspects of physical reality. These could in principle be measured, but may not be for practical reasons. In this situation, the term ''hidden variables'' is commonly used, (reflecting the fact that the variables are meaningful, but not observable). Other latent variables correspond to abstract concepts, like categories, behavioral or mental states, or data structures. The terms ''hypothetical variables'' or ''hypothetical constructs'' may be used in these situations.

The use of latent variables can serve to [[Dimensionality reduction|reduce the dimensionality]] of data. Many observable variables can be aggregated in a model to represent an underlying concept, making it easier to understand the data. In this sense, they serve a function similar to that of scientific theories. At the same time, latent variables link observable "[[sub-symbolic]]" data in the real world to symbolic data in the modeled world.

Line 7 ⟶ 13:

==Examples ==

[[File:Estimation of a mean height curve for boys from the Berkeley Growth Study with and without warping.gif|thumb|upright=2|Estimation of a mean height curve (black) for boys from the Berkeley Growth Study with and without warping. The warping is based on latent variables that maps age to a synchronized biological age using a [[nonlinear mixed-effects model]].<ref name="Raket_et_al_2014">{{cite journal |vauthors=Raket LL, Sommer S, Markussen B |year=2014 |title=A nonlinear mixed-effects model for simultaneous smoothing and registration of functional data |journal=Pattern Recognition Letters |volume=38|pages=1-71–7 |doi=10.1016/j.patrec.2013.10.018|bibcode=2014PaReL..38....1R }}</ref>]]

===Psychology===

Latent variables, as created by factor analytic methods, generally represent "shared" variance, or the degree to which variables "move" together. Variables that have no correlation cannot result in a latent construct based on the common [[Factor analysis|factor model]].<ref>{{cite book |last1=Tabachnick |first1=B.G. |last2=Fidell |first2=L.S. |title=Using Multivariate Analysis |publisher=Allyn and Bacon |___location=Boston |year=2001 |isbn=978-0-321-05677-1 }}{{Page needed|date=November 2010}}</ref>

Line 16 ⟶ 22:

* extraversion<ref name="status">{{cite journal |last1=Borsboom |first1=D. |author2-link=Gideon J. Mellenbergh |last2=Mellenbergh, G.J. |last3=van Heerden |first3=J. |title=The Theoretical Status of Latent Variables |journal=Psychological Review |volume=110 |issue=2 |pages=203–219 |year=2003 |doi=10.1037/0033-295X.110.2.203 |pmid=12747522 |url=http://rhowell.ba.ttu.edu/BorsboomLatentvars2003.pdf |citeseerx=10.1.1.134.9704 |access-date=2008-04-08 |archive-url=https://web.archive.org/web/20130120044039/http://rhowell.ba.ttu.edu/BorsboomLatentvars2003.pdf |archive-date=2013-01-20 |url-status=dead }}</ref>

* spatial ability<ref name="status"/>

* wisdom: “Two of the more predominant means of assessing wisdom include wisdom-related performance and latent variable measures.”<ref name="wisdom">{{cite journal |last1=Greene |first1=Jeffrey A. |last2=Brown |first2=Scott C. |title=The Wisdom Development Scale: Further Validity Investigations |journal=International Journal of Aging and Human Development |volume=68 |issue=4 |pages=289–320 (at p. 291) |year=2009 |pmid=19711618 |doi=10.2190/AG.68.4.b }}</ref>

* [[Spearman's g]], or the [[g factor (psychometrics)|general intelligence factor]] in [[psychometrics]]<ref>{{Cite journal | last1 = Spearman | first1 = C. | author-link = Charles Spearman| title = "General Intelligence," Objectively Determined and Measured | journal = The American Journal of Psychology | volume = 15 | issue = 2 | pages = 201–292 | doi = 10.2307/1412107 | year = 1904 | jstor = 1412107 }}</ref>

===Economics===

{{unreferenced section|date=January 2024}}

Examples of latent variables from the field of [[economics]] include [[quality of life]], business confidence, morale, happiness and conservatism: these are all variables which cannot be measured directly. ButHowever, by linking these latent variables to other, observable variables, the values of the latent variables can be inferred from measurements of the observable variables. Quality of life is a latent variable which cannot be measured directly, so observable variables are used to infer quality of life. Observable variables to measure quality of life include wealth, employment, environment, physical and mental health, education, recreation and leisure time, and social belonging.

===Medicine===

{{unreferenced section|date=January 2024}}

Latent-variable methodology is used in many branches of [[medicine]]. A class of problems that naturally lend themselves to latent variables approaches are [[longitudinal studies]] where the time scale (e.g. age of participant or time since study baseline) is not synchronized with the trait being studied. For such studies, an unobserved time scale that is synchronized with the trait being studied can be modeled as a transformation of the observed time scale using latent variables. Examples of this include [[Nonlinear_mixed-effects_model#Example: Disease progression modeling|disease progression modeling]] and [[Nonlinear_mixed-effects_model#Example: Growth analysis|modeling of growth]] (see box).

Line 34 ⟶ 40:

* [[Hidden Markov model]]s

* [[Factor analysis]]

* [[Item response theory]]

Analysis and inference methods include:

* [[Principal component analysis]]

* [[Instrumented principal component analysis]]<ref>Kelly, Bryan T. and Pruitt, Seth and Su, Yinan, Instrumented Principal Component Analysis (December 17, 2020). Available at SSRN: https://papers.ssrn.com/sol3abstract=2983919 or http://papersdx.doi.org/10.2139/ssrn.cfm?abstract_id=2983919</ref>

* [[Partial least squares regression]]

* [[Latent semantic analysis]] and [[probabilistic latent semantic analysis]]

Line 44 ⟶ 51:

===Bayesian algorithms and methods===

{{unreferenced section|date=January 2024}}

[[Bayesian statistics]] is often used for inferring latent variables.

* [[Latent Dirichlet allocation]]

* The [[Chinese restaurant process]] is often used to provide a prior distribution over assignments of objects to latent categories.

* The [[Indian buffet process]] is often used to provide a prior distribution over assignments of latent binary features to objects.

==See also==

Line 55 ⟶ 62:

* [[Confounding]]

* [[Dependent and independent variables]]

* [[Errors-in-variables models]]

* [[Evidence lower bound]]

* [[Factor analysis]]

* [[Intervening variable]]

* [[Latent variable model]]