Alogistic function orlogistic curve is a common S-shaped curve (sigmoid curve) with the equation
where
The logistic function has domain thereal numbers, the limit as is 0, and the limit as is.
Theexponential function with negated argument() is used to define thestandard logistic function where, which has the equationand is sometimes simply called thesigmoid function.[2] It is also sometimes called theexpit, being the inverse function of thelogit.[3][4]
The logistic function finds applications in a range of fields, includingbiology (especiallyecology),biomathematics,chemistry,demography,economics,geoscience,mathematical psychology,probability,sociology,political science,linguistics,statistics, andartificial neural networks. There are variousgeneralizations, depending on the field.
The logistic function was introduced in a series of three papers byPierre François Verhulst between 1838 and 1847, who devised it as a model ofpopulation growth by adjusting theexponential growth model, under the guidance ofAdolphe Quetelet.[5] Verhulst first devised the function in the mid 1830s, publishing a brief note in 1838,[1] then presented an expanded analysis and named the function in 1844 (published 1845);[a][6] the third paper adjusted the correction term in his model of Belgian population growth.[7]
The initial stage of growth is approximately exponential (geometric); then, as saturation begins, the growth slows to linear (arithmetic), and at maturity, growth approaches the limit with an exponentially decaying gap, like the initial stage in reverse.
Verhulst did not explain the choice of the term "logistic" (French:logistique), but it is presumably in contrast to thelogarithmic curve,[8][b] and by analogy with arithmetic and geometric. His growth model is preceded by a discussion ofarithmetic growth andgeometric growth (whose curve he calls alogarithmic curve, instead of the modern termexponential curve), and thus "logistic growth" is presumably named by analogy,logistic being fromAncient Greek:λογιστικός,romanized: logistikós, a traditional division ofGreek mathematics.[c]
As a word derived from ancient Greek mathematical terms,[9]the name of this function is unrelated to the military and management termlogistics, which is instead fromFrench:logis "lodgings",[10] though some believe the Greek term also influencedlogistics;[9] seeLogistics § Origin for details.
Thestandard logistic function is the logistic function with parameters,,, which yields
In practice, due to the nature of theexponential function, it is often sufficient to compute the standard logistic function for over a small range of real numbers, such as a range contained in [−6, +6], as it quickly converges very close to its saturation values of 0 and 1.
The logistic function has the symmetry property that
This reflects that the growth from 0 when is small is symmetric with the decay of the gap to the limit (1) when is large.
Further, is anodd function.
The sum of the logistic function and its reflection about the vertical axis,, is
The logistic function is thus rotationally symmetrical about the point (0, 1/2).[11]
The logistic function is the inverse of the naturallogit function
and so converts the logarithm ofodds into aprobability. The conversion from thelog-likelihood ratio of two alternatives also takes the form of a logistic curve.
The logistic function is an offset and scaledhyperbolic tangent function:or
This follows from
The hyperbolic-tangent relationship leads to another form for the logistic function's derivative:
which ties the logistic function into thelogistic distribution.
Geometrically, the hyperbolic tangent function is thehyperbolic angle on theunit hyperbola, which factors as, and thus has asymptotes the lines through the origin with slope and with slope, and vertex at corresponding to the range and midpoint () of tanh. Analogously, the logistic function can be viewed as the hyperbolic angle on the hyperbola, which factors as, and thus has asymptotes the lines through the origin with slope and with slope, and vertex at, corresponding to the range and midpoint () of the logistic function.
Parametrically,hyperbolic cosine andhyperbolic sine give coordinates on the unit hyperbola:[d], with quotient the hyperbolic tangent. Similarly, parametrizes the hyperbola, with quotient the logistic function. These correspond tolinear transformations (and rescaling the parametrization) ofthe hyperbola, with parametrization: the parametrization of the hyperbola for the logistic function corresponds to and the linear transformation, while the parametrization of the unit hyperbola (for the hyperbolic tangent) corresponds to the linear transformation.
The standard logistic function has an easily calculatedderivative. The derivative is known as the density of thelogistic distribution:
from which all higher derivatives can be derived algebraically. For example,.
The logistic distribution is alocation–scale family, which corresponds to parameters of the logistic function. If is fixed, then the midpoint is the location and the slope is the scale.
Conversely, itsantiderivative can be computed by thesubstitution, since
so (dropping theconstant of integration)
Inartificial neural networks, this is known as thesoftplus function and (with scaling) is a smooth approximation of theramp function, just as the logistic function (with scaling) is a smooth approximation of theHeaviside step function.
The standard logistic function isanalytic on the whole real line since, where, and, are analytic on their domains, and the composition of analytic functions is again analytic.
A formula for thenth derivative of the standard logistic function is
therefore itsTaylor series about the point is
The unique standard logistic function is the solution of the simple first-order non-linearordinary differential equation
withboundary condition. This equation is the continuous version of thelogistic map. Note that the reciprocal logistic function is solution to a simple first-orderlinear ordinary differential equation.[12]
The qualitative behavior is easily understood in terms of thephase line: the derivative is 0 when the function is 1; and the derivative is positive for between 0 and 1, and negative for above 1 or less than 0 (though negative populations do not generally accord with a physical model). This yields an unstable equilibrium at 0 and a stable equilibrium at 1, and thus for any function value greater than 0 and less than 1, it grows to 1.
The logistic equation is a special case of theBernoulli differential equation and has the following solution:
Choosing the constant of integration gives the other well known form of the definition of the logistic curve:
More quantitatively, as can be seen from the analytical solution, the logistic curve shows earlyexponential growth for negative argument, which reaches to linear growth of slope 1/4 for an argument near 0, then approaches 1 with an exponentially decaying gap.
The differential equation derived above is a special case of a general differential equation that only models the sigmoid function for. In many modeling applications, the moregeneral form[citation needed] can be desirable. Its solution is the shifted and scaledsigmoid function.
When the capacity, the value of the logistic function is in the range and can be interpreted as a probabilityp.[e] In more detail,p can be interpreted as the probability of one of two alternatives (the parameter of aBernoulli distribution);[f] the two alternatives are complementary, so the probability of the other alternative is and. The two alternatives are coded as 1 and 0, corresponding to the limiting values as.
In this interpretation the inputx is thelog-odds for the first alternative (relative to the second, measured in "logistic units" orlogits), and so is theodds for the first alternative (relative to the second). Given odds for an event of ( against1), the probability is the ratio of "for" over "for plus against",. We see that the logistic function,, is the probability of the first alternative.
Conversely,x is the log-oddsagainst the second alternative, is the log-oddsfor the second alternative, is the odds for the second alternative, and is the probability of the second alternative.
This can be framed more symmetrically in terms of two inputs, and, which then generalizes naturally to more than two alternatives. Given two real number inputs, and, interpreted as logits, theirdifference is the log-odds for option 1 (the log-oddsagainst option 0), is the odds, is the probability of option 1, and similarly is the probability of option 0.
This form immediately generalizes to more alternatives as thesoftmax function, which is a vector-valued function whosei-th coordinate is.
More subtly, the symmetric form emphasizes interpreting the inputx as and thusrelative to some reference point, implicitly to. Notably, the softmax function is invariant under adding a constant to all the logits, which corresponds to the difference being the log-odds for optionj against optioni, but the individual logits not being log-odds on their own. Often one of the options is used as a reference ("pivot"), and its value fixed as0, so the other logits are interpreted as odds versus this reference. This is generally done with the first alternative, hence the choice of numbering:, and then is the log-odds for optioni against option0. Since, this yields the term in many expressions for the logistic function and generalizations.[g]
In growth modeling, numerous generalizations exist, including thegeneralized logistic curve, theGompertz function, thecumulative distribution function of theshifted Gompertz distribution, and thehyperbolastic function of type I.
In statistics, where the logistic function is interpreted as the probability of one of two alternatives, the generalization to three or more alternatives is thesoftmax function, which is vector-valued, as it gives the probability of each alternative.
A typical application of the logistic equation is a common model ofpopulation growth (see alsopopulation dynamics), originally due toPierre-François Verhulst in 1838, where the rate of reproduction is proportional to both the existing population and the amount of available resources, all else being equal. The Verhulst equation was published after Verhulst had readThomas Malthus'An Essay on the Principle of Population, which describes theMalthusian growth model of simple (unconstrained) exponential growth. Verhulst derived his logistic equation to describe the self-limiting growth of abiological population. The equation was rediscovered in 1911 byA. G. McKendrick for the growth of bacteria in broth and experimentally tested using a technique for nonlinear parameter estimation.[13] The equation is also sometimes called theVerhulst-Pearl equation following its rediscovery in 1920 byRaymond Pearl (1879–1940) andLowell Reed (1888–1966) of theJohns Hopkins University.[14] Another scientist,Alfred J. Lotka derived the equation again in 1925, calling it thelaw of population growth.
Letting represent population size ( is often used in ecology instead) and represent time, this model is formalized by thedifferential equation:
where the constant defines thegrowth rate and is thecarrying capacity.
In the equation, the early, unimpeded growth rate is modeled by the first term. The value of the rate represents the proportional increase of the population in one unit of time. Later, as the population grows, the modulus of the second term (which multiplied out is) becomes almost as large as the first, as some members of the population interfere with each other by competing for some critical resource, such as food or living space. This antagonistic effect is called thebottleneck, and is modeled by the value of the parameter. The competition diminishes the combined growth rate, until the value of ceases to grow (this is calledmaturity of the population).The solution to the equation (with being the initial population) is
where
where is the limiting value of, the highest value that the population can reach given infinite time (or come close to reaching in finite time). The carrying capacity is asymptotically reached independently of the initial value, and also in the case that.
In ecology,species are sometimes referred to as-strategist or-strategist depending upon theselective processes that have shaped theirlife history strategies.Choosing the variable dimensions so that measures the population in units of carrying capacity, and measures time in units of, gives the dimensionless differential equation
Theantiderivative of the ecological form of the logistic function can be computed by thesubstitution, since
Since the environmental conditions influence the carrying capacity, as a consequence it can be time-varying, with, leading to the following mathematical model:
A particularly important case is that of carrying capacity that varies periodically with period:
It can be shown[15] that in such a case, independently from the initial value, will tend to a unique periodic solution, whose period is.
A typical value of is one year: In such case may reflect periodical variations of weather conditions.
Another interesting generalization is to consider that the carrying capacity is a function of the population at an earlier time, capturing a delay in the way population modifies its environment. This leads to a logistic delay equation,[16] which has a very rich behavior, with bistability in some parameter range, as well as a monotonic decay to zero, smooth exponential growth, punctuated unlimited growth (i.e., multiple S-shapes), punctuated growth or alternation to a stationary level, oscillatory approach to a stationary level, sustainable oscillations, finite-time singularities as well as finite-time death.
Logistic functions are used in several roles in statistics. For example, they are thecumulative distribution function of thelogistic family of distributions, and they are, a bit simplified, used to model the chance a chess player has to beat their opponent in theElo rating system. More specific examples now follow.
Logistic functions are used inlogistic regression to model how the probability of an event may be affected by one or moreexplanatory variables: an example would be to have the model
where is the explanatory variable, and are model parameters to be fitted, and is the standard logistic function.
Logistic regression and otherlog-linear models are also commonly used inmachine learning. A generalisation of the logistic function to multiple inputs is thesoftmax activation function, used inmultinomial logistic regression.
Another application of the logistic function is in theRasch model, used initem response theory. In particular, the Rasch model forms a basis formaximum likelihood estimation of the locations of objects or persons on acontinuum, based on collections ofcategorical data, for example the abilities of persons on a continuum based on responses that have been categorized as correct and incorrect.
Logistic functions are often used inartificial neural networks to introducenonlinearity in the model or to clamp signals to within a specifiedinterval. A popularneural net element computes alinear combination of its input signals, and applies a bounded logistic function as theactivation function to the result; this model can be seen as a "smoothed" variant of the classicalthreshold neuron.
A common choice for the activation or "squashing" functions, used to clip large magnitudes to keep the response of the neural network bounded,[17] is
which is a logistic function.
These relationships result in simplified implementations ofartificial neural networks withartificial neurons. Practitioners caution that sigmoidal functions which areantisymmetric about the origin (e.g. thehyperbolic tangent) lead to faster convergence when training networks withbackpropagation.[18]
The logistic function is itself the derivative of another proposed activation function, thesoftplus.
Another application of logistic curve is in medicine, where the logistic differential equation can be used to model the growth oftumors. This application can be considered an extension of the above-mentioned use in the framework of ecology (see also theGeneralized logistic curve, allowing for more parameters). Denoting with the size of the tumor at time, its dynamics are governed by
which is of the type
where is the proliferation rate of the tumor.
If a course ofchemotherapy is started with a log-kill effect, the equation may be revised to be
where is the therapy-induced death rate. In the idealized case of very long therapy, can be modeled as a periodic function (of period) or (in case of continuous infusion therapy) as a constant function, and one has that
i.e. if the average therapy-induced death rate is greater than the baseline proliferation rate, then there is the eradication of the disease. Of course, this is an oversimplified model of both the growth and the therapy. For example, it does not take into account the evolution of clonal resistance, or the side-effects of the therapy on the patient. These factors can result in the eventual failure of chemotherapy, or its discontinuation.[citation needed]
A novel infectious pathogen to which a population has no immunity will generally spread exponentially in the early stages, while the supply of susceptible individuals is plentiful. The SARS-CoV-2 virus that causesCOVID-19 exhibited exponential growth early in the course of infection in several countries in early 2020.[19] Factors including a lack of susceptible hosts (through the continued spread of infection until it passes the threshold forherd immunity) or reduction in the accessibility of potential hosts through physical distancing measures, may result in exponential-looking epidemic curves first linearizing (replicating the "logarithmic" to "logistic" transition first noted byPierre-François Verhulst, as noted above) and then reaching a maximal limit.[20]
A logistic function, or related functions (e.g. theGompertz function) are usually used in a descriptive or phenomenological manner because they fit well not only to the early exponential rise, but to the eventual levelling off of the pandemic as the population develops a herd immunity. This is in contrast to actual models of pandemics which attempt to formulate a description based on the dynamics of the pandemic (e.g. contact rates, incubation times, social distancing, etc.). Some simple models have been developed, however, which yield a logistic solution.[21][22][23]
Ageneralized logistic function, also called the Richards growth curve, has been applied to model the early phase of theCOVID-19 outbreak.[24] The authors fit the generalized logistic function to the cumulative number of infected cases, here referred to asinfection trajectory. There are different parameterizations of thegeneralized logistic function in the literature. One frequently used forms is
where are real numbers, and is a positive real number. The flexibility of the curve is due to the parameter: (i) if then the curve reduces to the logistic function, and (ii) as approaches zero, the curve converges to theGompertz function. In epidemiological modeling,,, and represent the final epidemic size, infection rate, and lag phase, respectively. See the right panel for an example infection trajectory when is set to.
One of the benefits of using a growth function such as thegeneralized logistic function in epidemiological modeling is its relatively easy application to themultilevel model framework, where information from different geographic regions can be pooled together.
The concentration of reactants and products inautocatalytic reactions follow the logistic function.The degradation ofPlatinum group metal-free (PGM-free) oxygen reduction reaction (ORR) catalyst in fuel cell cathodes follows the logistic decay function,[25] suggesting an autocatalytic degradation mechanism.
The logistic function determines the statistical distribution of fermions over the energy states of a system in thermal equilibrium. In particular, it is the distribution of the probabilities that each possible energy level is occupied by a fermion, according toFermi–Dirac statistics.
The logistic function also finds applications in optics, particularly in modelling phenomena such asmirages. Under certain conditions, such as the presence of a temperature or concentration gradient due to diffusion and balancing with gravity, logistic curve behaviours can emerge.[26][27]
A mirage, resulting from a temperature gradient that modifies the refractive index related to the density/concentration of the material over distance, can be modelled using a fluid with a refractive index gradient due to the concentration gradient. This mechanism can be equated to a limiting population growth model, where the concentrated region attempts to diffuse into the lower concentration region, while seeking equilibrium with gravity, thus yielding a logistic function curve.[26]
In linguistics, the logistic function can be used to modellanguage change:[28] an innovation that is at first marginal begins to spread more quickly with time, and then more slowly as it becomes more universally adopted.
The logistic S-curve can be used for modeling the crop response to changes in growth factors. There are two types of response functions:positive andnegative growth curves. For example, the crop yield mayincrease with increasing value of the growth factor up to a certain level (positive function), or it maydecrease with increasing growth factor values (negative function owing to a negative growth factor), which situation requires aninverted S-curve.
The logistic function can be used to illustrate the progress of thediffusion of an innovation through its life cycle.
InThe Laws of Imitation (1890),Gabriel Tarde describes the rise and spread of new ideas through imitative chains. In particular, Tarde identifies three main stages through which innovations spread: the first one corresponds to the difficult beginnings, during which the idea has to struggle within a hostile environment full of opposing habits and beliefs; the second one corresponds to the properly exponential take-off of the idea, with; finally, the third stage is logarithmic, with, and corresponds to the time when the impulse of the idea gradually slows down while, simultaneously new opponent ideas appear. The ensuing situation halts or stabilizes the progress of the innovation, which approaches an asymptote.
In asovereign state, the subnational units (constituent states or cities) may use loans to finance their projects. However, this funding source is usually subject to strict legal rules as well as to economyscarcity constraints, especially the resources the banks can lend (due to theirequity orBasel limits). These restrictions, which represent a saturation level, along with an exponential rush in aneconomic competition for money, create apublic finance diffusion of credit pleas and the aggregate national response is asigmoid curve.[31]
Historically, when new products are introduced there is an intense amount ofresearch and development which leads to dramatic improvements in quality and reductions in cost. This leads to a period of rapid industry growth. Some of the more famous examples are: railroads, incandescent light bulbs,electrification, cars and air travel. Eventually, dramatic improvement and cost reduction opportunities are exhausted, the product or process are in widespread use with few remaining potential new customers, and markets become saturated.
Logistic analysis was used in papers by several researchers at the International Institute of Applied Systems Analysis (IIASA). These papers deal with the diffusion of various innovations, infrastructures and energy source substitutions and the role of work in the economy as well as with the long economic cycle. Long economic cycles were investigated by Robert Ayres (1989).[32] Cesare Marchetti published onlong economic cycles and on diffusion of innovations.[33][34] Arnulf Grübler's book (1990) gives a detailed account of the diffusion of infrastructures including canals, railroads, highways and airlines, showing that their diffusion followed logistic shaped curves.[35]
Carlota Perez used a logistic curve to illustrate the long (Kondratiev) business cycle with the following labels: beginning of a technological era asirruption, the ascent asfrenzy, the rapid build out assynergy and the completion asmaturity.[36]
Logistic growth regressions carry significant uncertainty when data is available only up to around the inflection point of the growth process. Under these conditions, estimating the height at which the inflection point will occur may have uncertainties comparable to the carrying capacity (K) of the system.
A method to mitigate this uncertainty involves using the carrying capacity from a surrogate logistic growth process as a reference point.[37] By incorporating this constraint, even if K is only an estimate within a factor of two, the regression is stabilized, which improves accuracy and reduces uncertainty in the prediction parameters. This approach can be applied in fields such as economics and biology, where analogous surrogate systems or populations are available to inform the analysis.
Link[38] created an extension ofWald's theory of sequential analysis to a distribution-free accumulation of random variables until either a positive or negative bound is first equaled or exceeded. Link[39] derives the probability of first equaling or exceeding the positive boundary as, the logistic function. This is the first proof that the logistic function may have a stochastic process as its basis. Link[40] provides a century of examples of "logistic" experimental results and a newly derived relation between this probability and the time of absorption at the boundaries.
Nous donnerons le nom delogistique à la courbe [We will give the namelogistic to the curve]
The diagram clinched it for me: there two curves labeled "Logistique" and "Logarithmique" are drawn on the same axes, and one can see that there is a region where they match almost exactly, and then diverge.
I concluded that Verhulst's intention in naming the curve was indeed to suggest this comparison, and that "logistic" was meant to convey the curve's "log-like" quality.
{{cite journal}}: CS1 maint: numeric names: authors list (link)