Let be independent and identically distributed random variables with expected value zero, variance one, and a moment-generating function finite in a neighborhood of zero. Each row of deterministic real weights is finite. If and , then the weighted sums converge in distribution to a normal distribution with mean zero and variance ; denotes the constant zero.
By the quadratic expansion of a cumulant-generating function, write with . For a fixed real , put , with the supremum taken as zero if its range is empty. Then and
Thus the log moment-generating function of the sum tends to . The convergence theorem for moment-generating functions proves the conclusion. The supremum is the step that makes the remainder estimate uniform across the row; a separate little-o assertion for each term would not justify summing an increasing number of terms.
For observations of an Itô integral on a partition of an interval, the statistic estimates . With a deterministic noise coefficient the squared increments are independent, and the fluctuation has variance .
A heteroscedastic extension of independent-groups one-way analysis of variance, allowing unequal group variances. It does not make paired or clustered observations independent.
Welch's t-test Created 2026-10-18 Updated 2026-10-25
For two independent samples from approximately normal distributions, Welch's t-test compares their means without assuming equal unknown variances. The displayed test statistic uses sample variances . Under the equal-mean null hypothesis, its reference Student t-distribution has approximate statistical degrees of freedom
A p-value from this approximation addresses the specified statistical error model; it cannot distinguish a material effect from an unmeasured experimenter bias when sample and experimenter are confounded.
White noise Created 2026-09-28 Updated 2026-10-14
For the AR(1) and MA(1) formulas here, white noise means a real sequence indexed by all integer times with
These are the weak white noise assumptions: zero mean, a common finite positive variance, and uncorrelated random variables at distinct times. Independence and a normal distribution are additional assumptions. Strong white noise requires independent and identically distributed random variables; discrete Gaussian white noise is the jointly Gaussian special case. Some definitions allow a constant nonzero mean; subtract it before using the centered formulas here.
For a concrete dependent example, let equal or with equal probabilities, independently of a sequence of independent Rademacher random variables , and put . Its mean is zero, its variance is , and its nonzero-lag autocovariances vanish. Nevertheless every equals the same : seeing one magnitude determines all the others. This is white noise with dependence invisible to its ACF, and is also an example of the issue discussed in ergodic stationary process.
White-noise likelihood for a square-integrable shift Created 2026-10-05 Updated 2026-10-06
Generalized Gaussian white noise over a real Hilbert space is an isonormal Gaussian process with covariance . Its law is carried on a larger observation space; the identity is not the covariance of an infinite-dimensional -valued Gaussian measure because it is not a trace-class operator. A deterministic signal defines a translated observation law with Radon-Nikodym derivative
In an orthonormal basis, is a stochastic series with total variance , not an inner product of two -valued observations. The formula follows from the Cameron-Martin theorem for a Gaussian measure in its generalized white-noise version, or from finite-dimensional Gaussian likelihood ratios. On , generalized noise can be realized on tempered distributions; an unweighted global negative Sobolev space is not automatically a suitable almost-sure support on an unbounded domain. The stochastic series avoids imposing that unsupported regularity.
Wigner matrix Created 2026-10-07 Updated 2026-11-05
A real Wigner matrix is a real symmetric matrix whose upper off-diagonal entries are independent centered copies of a fixed random variable, scaled to have variance of order . Diagonal conventions may vary and should be specified separately. A common version has variance-one unscaled off-diagonal entries and either zero diagonal entries or independent diagonal entries with a bounded second moment.
Wilcoxon signed-rank statistic Created 2026-10-14 Updated 2026-11-05
For a continuous sample, the statistic also equals . In each pair the sign of the sum is the sign of the observation with larger absolute value, proving the identity by counting its smaller-rank partners. Under a continuous symmetric null, the pair kernel has projection and projection variance . The U-statistic central limit theorem yields asymptotic normality of , with variance asymptotic to .
Wild bootstrap 2026-10-14
For independent mean-zero errors in a fixed-design regression, form , where the independent multipliers have mean zero and variance one, and refit. Rademacher multipliers taking the values and with equal probability are a simple choice. Conditional on the data, the bootstrap errors have mean zero and preserve observation-specific residual variances, supporting heteroscedasticity rather than imposing a common residual distribution. Appropriate leverage corrections and regularity assumptions are required for accurate studentized inference.
William Sealy Gosset Created 2026-10-18 Updated 2026-11-05
William Sealy Gosset published statistical research under the name Student while working as a brewer for Guinness. His practical problems involved variation in ingredients and brewing conditions, where obtaining a large sample was often difficult. This motivated methods that account for uncertainty in estimating variability from a small sample.[ref]
The Student's t-distribution describes the standardized mean of an independent normal sample when its unknown standard deviation is replaced by the sample standard deviation. Its heavier tails, compared with a Standard normal distribution, reflect that extra uncertainty. Gosset's work helped turn the limitations of a small experiment into an explicit part of inference rather than treating a variance estimate as known.
Figure 1. . Unknown photographer; scan credited to Wujaszek; subsequent descreening by Quibik. Public domain; Public Domain Mark 1.0. Source and usage information on the linked Wikimedia file page. Source.
WinBUGS 2026-10-14
Software for Bayesian inference through Markov chain Monte Carlo in directed statistical models. Its model language uses dnorm(mean, precision), so the second argument is inverse variance; dcat indexes a supplied probability vector, and dbern introduces a binary latent variable. The model is declarative, so the order of deterministic node definitions does not impose procedural evaluation order.
Wind-tunnel turbulence intensity describes velocity fluctuations in a stated test region relative to its mean flow. For the streamwise velocity , define and . The streamwise turbulence intensity is
Its numerator is the root mean square of the fluctuation, equivalently its standard deviation when the same averaging convention is used. It is not the raw root mean square of , which still contains the mean. Component direction, averaging duration, probe position and retained frequency band belong with the reported value.
A hot-wire anemometer converts changes in convective cooling into an electrical signal; measurement calibration relates that signal to speed. Dan H. Neuhart and Catherine B. McGinley's NASA study used a survey rake and additional probes, including pressure and temperature measurements for hot-wire calibration. It separated mean and fluctuating channels and used reference wires to identify coherent disturbances. Its frequency-band comparisons show why filtering can change a reported intensity.[ref]
For an original independent-noise example, suppose measured fluctuations have standard deviation , additive sensor noise has , and . If their covariance vanishes, the physical fluctuation variance is , giving , rather than the uncorrected . Correlated noise requires an additional covariance term; an estimated noise variance larger than the measured variance signals an inconsistent subtraction model.
Slow drift is another distinct contribution. A deterministic mean change over an interval of length has time-averaged variance . With and , its apparent fluctuation standard deviation is , despite no random turbulence in this example. This mean-drift contribution to a turbulence-intensity estimate cannot be distinguished from turbulence by a single unqualified RMS number.
A tunnel survey should therefore retain the signal-processing convention and spatial positions, rather than only a facility-wide percentage. Changes in turbulence intensity can affect boundary-layer transition; equal Reynolds numbers do not guarantee equal fluctuation environments.
Consider an original idealized track reconstruction model. Let equally spaced wires have positions and assign each event to the nearest wire, reporting . Assume exactly one correct wire address, perfect detection efficiency, exact wire positions and a uniform distribution of true positions within each cell. These assumptions explicitly exclude missed pulses, false addresses and several unresolved tracks.
Writing , the measurement error is , with uniform on . Symmetry gives zero expected value. Its variance and standard deviation follow by integration:
For chosen wire pitches and this gives and respectively. These are model standard deviations, not measured chamber spatial resolutions. The largest possible absolute error in this model is , which is a different quantity.
A useful subtlety is that the formula depends on the illumination. For an unresolved pencil beam fixed at , every reconstructed position is . The repeated estimates have zero variance, but every error is . Small repeat-to-repeat spread can coexist with a biased position. The uniform-cell result therefore cannot be transferred to an arbitrary beam profile or detector selection.
The wire address supplies one transverse coordinate in this model. Combining several separated measurements supports track reconstruction; a single address does not specify a complete trajectory. If two charged particles enter the same cell during one accepted interval, an address-only record can be identical to that from one particle. No reconstruction of that record can recover the lost multiplicity without additional information.
Figure 1.
Chosen wire-address errors and a calibrated two-wire response model
. Original teaching calculations, not measured data. Left: uniform-cell address error. Right: stipulated linear charge response with independent small channel noise; the noise estimate is a first-order approximation, not a chamber performance claim. Details.
The drive can supply or extract work while the quantum state stays approximately fixed. For a differentiable bounded Hamiltonian operator, the Schrödinger equation gives
The state-evolution terms cancel, leaving the external drive's power. For a finite ideal step the change in mean energy is . Thus conservation of energy for a closed system with a time-independent Hamiltonian operator does not assert conservation of the driven subsystem's energy during switching.
Specify work operationally by two energy measurements: one before the switch and one after. For the prepared initial ground state, the first outcome is definite; the second samples the new Born rule probabilities. This protocol gives a distribution of measured energy differences, rather than generally the measurement distribution of the single operator .[ref]
and hence
For and , the work is or , with probabilities and ; its expectation is and its variance is . The initial mean energy becomes . Using different initial and final scalar offsets would also shift the measured work; the displayed Hamiltonian operators fix that convention explicitly.
A working correlation matrix is a positive-definite proposed correlation structure used to weight a generalized estimating equation, with the diagonal matrix of marginal variances. It need not equal the actual correlation matrix. Correct marginal means and independent sampling clusters keep the estimating equations unbiased; a cluster-level sandwich covariance matrix estimates their actual large-sample covariance. For three exchangeable observations, positive definiteness requires .
Wright path tracing rule Created 2026-09-24 Updated 2026-11-05
Wright path tracing rule computes covariances in a linear structural equation model by adding contributions along the admissible paths of its graph. For an acyclic system with centered disturbances and disturbance covariance matrix ,
Acyclicity makes nilpotent in a topological ordering, so . Every matrix entry sums products of coefficients along directed paths. Substituting this finite expansion into the covariance formula gives the path-tracing construction. With uncorrelated disturbances, contributions join two directed paths at a common disturbance; correlated disturbances add the corresponding cross-disturbance terms.
For a worked example, let , and , with mutually uncorrelated centered disturbances. Substitution gives , hence
The direct path contributes and the mediated path contributes , each times the source variance. For standardized variables this becomes a sum of coefficient products for a correlation coefficient. Reading those coefficients as causal effects requires the structural intervention interpretation; an arbitrary covariance-fitting graph does not supply it by itself.
Wringing-allocation pilot uncertainty is uncertainty in the noise components used to choose how many gauge block contacts to prepare and how many readings to take per contact. The optimum in nested contact-study allocation depends on estimated between-contact and within-contact variability; a short pilot does not determine their ratio exactly.
For a balanced pilot with independent wrings and readings per wring, use the random-intercept linear mixed model
with all random effects independent. The within-wring mean square has expected value ; the between-wring mean square has expected value . Thus method-of-moments planning estimates are
Truncating a negative estimate is a boundary convention, not proof that true between-wring variance vanishes. The expectations follow by decomposing the independent within-wring residuals and wring means in this stated model.
An original pilot has , and . The estimates are nm and nm. If another short pilot gives with the same , truncation instead gives . Substituting that boundary value into the continuous planning rule drives the preferred readings per wring toward infinity, revealing a fragile design decision rather than an experimentally established absence of contact variability.
With independent normal residuals, has relative standard deviation , where . Five wrings with two readings each give and approximately 63% relative scatter in this variance estimate. The between-component estimate subtracts two noisy mean squares and can be still less stable. More readings of the same contact do not substitute for observing more independent contacts.
A useful planning sensitivity analysis holds readout standard deviation at a stipulated 6 nm while testing plausible from 1 to 4 nm. With preparation cost 9 minutes and reading cost 1 minute, the continuous optimum
ranges from 18 to 4.5 readings per wring. Under a 180-minute budget, feasible allocations give mean variances approximately when nm, but when nm. The best choice among these candidates changes with the pilot assumption; no one set of counts is universally optimal.
Cluster-resampled wringing allocation can retain the pilot's contact grouping when assessing planning sensitivity; treating all readings as independent destroys the nesting. Any resampling still inherits the pilot's limited contact population. Physically, record fresh preparation, waiting time and handling: NIST documents changes in wringing films with conditions and elapsed time.[ref] Drift and common reference errors require their own model terms rather than being silently absorbed into independent readout noise.
Yule-Walker equations 2026-10-14
For a causal autoregressive model with centered white noise of variance , multiplying by past values and taking expectations yields for , and . These Yule-Walker equations determine the autocovariance from the coefficients and noise variance, or estimate coefficients from observed autocovariances. Causality supplies the needed orthogonality of the current noise to past observations.
Zero-energy filament mode Created 2026-10-07 Updated 2026-10-14
A zero-energy mode lies in the kernel of the quadratic elastic filament operator. For pure bending, zero curvature makes the displacement affine. Free-free endpoints therefore leave translation and tilt modes, while slope-constrained, force-free endpoints leave only translation. Both clamped-clamped and hinged-hinged endpoints remove these modes. An unrestricted modal amplitude has constant Boltzmann distribution weight, so no normalizable canonical ensemble or finite displacement variance exists. Fixing the rigid degrees of freedom allows the equipartition theorem on the positive complement.
Zero-inflated negative binomial model Created 2026-10-05 Updated 2026-10-07
With structural-zero probability , size and count mean , the model assigns to zero and to positive counts. Its expectation is and its variance is , by the law of total variance. A logarithmic link function can relate the count-component mean to predictors.