For the expectation-maximization algorithm, let . At iteration , the E-step forms the conditional distribution of the missing data at the current parameter and the function
The M-step is
In this maximization the conditional distribution defining the expectation is held at the old parameter; it is not recomputed as varies. The Jensen inequality gives
by writing the likelihood function ratio as a conditional expectation of . Hence an exact M-step does not decrease the observed-data likelihood function. The algorithm does not in general guarantee convergence to its global maximum; initialization and the likelihood function's geometry matter.
For the specified diagonal bivariate normal distribution, use . The two coordinates are independent, as are the four observations. Therefore the known second coordinate of the fourth observation gives no information about its first coordinate conditional on the parameters. The E-step is
These two moments suffice for the complete Gaussian log-likelihood. Discarding terms independent of the proposed parameters, its conditional expectation is
Differentiating with respect to each mean gives four times that mean minus the corresponding expected sum of observations. Thus the updates in EM for an independent missing normal coordinate are
For either variance the remaining criterion is ; its derivative vanishes at , which is its maximum when . Evaluate the expected residual sum at the new means to obtain
The missing contribution is
Its first term is essential. Substituting only the imputed mean for would omit its uncertainty and would not maximize the required expected complete log-likelihood.
As a check, the observed-data likelihood function integrates out the missing first coordinate, leaving three first-coordinate and four second-coordinate normal densities. Put and . The first-coordinate recurrence implies
Thus the iterates converge to the three-observation mean and maximum-likelihood variance, while the second-coordinate estimates are reached after one update. If the observed values in a coordinate are all equal, its likelihood function instead has a zero-variance boundary degeneracy, and there is no positive-variance maximizer; the variance formulas then describe that boundary limit. For nonzero empirical spreads the displayed updates give the ordinary nonsingular EM iteration.
The Metropolis–Hastings algorithm proposes and moves from the current state with acceptance probability
On rejection it retains . The target need only be known up to its normalizing constant, but it must define a proper probability distribution. The accepted off-diagonal probability flow is
which is symmetric in and proves detailed balance. This permits considerable flexibility in proposals, including joint moves when a posterior distribution has strongly correlated parameters, but requires proposal tuning and can waste computation on rejections.
A Gibbs sampler instead updates a coordinate or block from its exact full conditional distribution, holding the other coordinates fixed. This is a special Metropolis–Hastings algorithm update with acceptance probability one: the joint density factors into the unchanged marginal density and the proposed conditional density, which cancel in the acceptance ratio. The coordinate update preserves the target; composing such updates therefore preserves it too, although an entire deterministic sweep need not itself be reversible.
Use Gibbs sampling when the conditional laws are easy to simulate and suitably chosen blocks mix well. Scalar updates can mix poorly under strong posterior distribution correlation, whereas a blocked Gibbs sampler can alleviate that dependence. Prefer a flexible Metropolis proposal when conditional draws are difficult or a well-tuned joint proposal is more effective. Variations include random or systematic coordinate scans, random-walk Metropolis algorithms, independence proposals, and using Metropolis updates for difficult blocks inside a Gibbs scheme. Both methods generate dependent draws; usable long-run estimates require an appropriate irreducible, aperiodic chain and assessment of mixing, not merely a formal update rule.
For the change-point calculation, put and
Multiplication of the Gaussian likelihood function by the specified prior distribution gives the formal posterior distribution kernel, with respect to and counting measure on ,
There is an empty component under an improper prior at . The sum is independent of , whose prior distribution is flat on all of . For every fixed finite and , the kernel is a strictly positive constant in , hence
The configuration has positive prior probability . Integrating over, for example, a bounded interval of and already makes the joint normalization infinite. Thus the printed prior distribution does not produce a posterior probability distribution. This failure is present for every data set, not just for a special arrangement of observations.
The formal full conditionals expose exactly where the requested sampler breaks. For , completing the two mean squares gives
The variance kernel is that of when , and the discrete conditional probabilities are proportional to . At the conditional kernel for is constant and cannot be normalized; it is not a normal distribution with an infinite variance. Therefore no valid Gibbs sampler, and no value of , exists for the model as printed. A chain that simply substitutes an arbitrary distribution for this missing conditional would be sampling a different model.
Here is a fully specified proper-prior repair retaining the no-change configuration. Choose fixed hyperparameters
independently, with . Its Gaussian change-point posterior with proper priors is
It is normalizable even for degenerate data: the likelihood function is at most , and its integral against the variance prior distribution is bounded by
The normal mean priors integrate to one, and the normalizing constant is positive. This proves posterior propriety for the repaired model.
Completing the conditional squares gives the Gaussian change-point Gibbs updates. Define
The four updates are
At , , giving its proper prior distribution unchanged. Starting from any finite means, positive variance and allowed , cycle through these four updates using the latest values. The sum-of-squares terms can be evaluated from cumulative data sums; normalize the discrete log weights after subtracting their maximum to avoid numerical underflow. The proper positive conditionals allow movement among all split values and throughout the continuous parameter support.
Under this specified repair, is the posterior probability of no change within the observed sequence: all data points belong to the first mean regime. For retained draws after initialization has ceased to affect the estimates, its Markov chain Monte Carlo estimator is
Alternatively average the displayed conditional probability of at the sampled continuous parameters, giving an estimate by Rao-Blackwellization. Autocorrelation must be accounted for in its Monte Carlo uncertainty. These are probabilities and estimators for the explicitly repaired model; the original unnormalizable kernel supplies neither.
An unordered bootstrap sample is determined by the multiplicities of the distinct observations, with . Conversely each such vector specifies exactly one sample up to rearrangement. Place identical marks into boxes using separators. The stars and bars argument therefore gives
These bootstrap count vectors are not equally likely: the probability of a vector is , counting its possible ordered sequences. The ordered index sequences themselves number .
The first simulation uses independent standard Cauchy distributions and returns
The tail probability and estimator moments are
The indicator count has a binomial distribution with parameters ; the infinite variance of a Cauchy observation is irrelevant because this estimator averages bounded indicators.
For the next code block, each row is an independent sample of size from the empirical distribution of the original . Its entry in the vector is the corresponding bootstrap indicator mean . The last expression is the variance of these replicate estimates, with denominator :
Condition on the original data and write a star on bootstrap expectations. Every resampled indicator has a Bernoulli distribution with success probability , and the draws within a row are independent. Thus the conditional bootstrap variance of a sample mean gives
It estimates the sampling variance of the tail-probability estimator, not the variance of the raw Cauchy data. Its square root is an estimated standard error. Finite leaves bootstrap simulation noise. Averaging over the original data, its expectation is , so it is not exactly unbiased for the true variance at finite .
In the importance-sampling block, the transformation gives, for ,
The variable is a proposal draw supported entirely in the integration region. The deterministic value computed from it is the importance sampling weight
Consequently
The code uses ordinary importance sampling of a Cauchy tail, not self-normalization; itself is neither an observation from the target Cauchy law nor the estimator's summand. The weight compensates for sampling from its proposal law.
The final block bootstraps these same proposal observations and recomputes the weight average in each row. Since is a deterministic function of , this is equivalent to resampling the observed weights . If denotes a replicate, the final output is its sample variance over the replicates. Its conditional expectation is
Its expectation over the original proposal sample is .
The variance reduction can be calculated, rather than inferred just from the appearance of the weights. Write . Then
because an antiderivative of the integrand without its prefactor is . Hence
By comparison, . For the true estimator variances are approximately and respectively. The importance estimator has about 1/1,317 of the indicator estimator's variance, because all proposals land in the tail and the weights remain in the narrow interval .
Thus the last bootstrap output should usually be much smaller than the first, and its expected value is smaller by the same variance ratio. This is not a deterministic ordering of two finite-run outputs: if all the original Cauchy draws happen to be at most two, the first bootstrap variance is zero, while unequal importance weights can still give a positive second variance.
Choose a proposal probability density function and a finite envelope constant such that wherever the target has positive density. A rejection sampling trial draws and, independently, ; accept when
otherwise repeat with fresh independent draws. For any measurable set ,
In particular , so the conditional accepted density is . More explicitly, summing over the possible first accepted trial gives
Thus the algorithm terminates almost surely and its output has exactly the requested law.
The envelope constant satisfies and is the expected number of proposal trials for one output, by the geometric distribution of the waiting time. This is the quantity often called the rejection rate . The actual probability of rejecting a particular proposal is , and the expected number of rejections before one acceptance is ; these conventions should not be conflated.
To generate a standard normal distribution, first generate its absolute value, whose half-normal distribution density is
With the given exponential distribution proposal, the density ratio is
Its maximum is at , hence the normal rejection sampling with an exponential envelope constant is
Draw and a fresh uniform ; accept when . An accepted has density . Independently choose a sign, positive or negative with probability , and output with that sign and magnitude . On either half-line its density is , proving the normal output law.
To optimize the rejection rate, differentiate
Its derivative is , negative before one and positive after one; equivalently its second derivative is . Thus
The rejection probability is consequently about .
To use only uniforms, apply inverse transform sampling: for an independent ,
has survival probability for . Use a second independent uniform for acceptance, and, after acceptance, a third independent uniform to choose the sign. Repeat with fresh uniforms after any rejection. No exponential generator is then needed.
Use the standard linear Gaussian state-space model convention: the initial state is independent of both mutually independent iid noise sequences. This independence from the initial state is needed for the usual forecast recursion. Let and write and . Taking conditional expectations in the state equation gives
To justify the covariance, the forecast error is . Its two terms have zero cross-covariance because is independent of the previous state and observations. Their covariances therefore add. In this Gaussian model the conditional error covariance depends on the parameters and time but not on the realized data, so it also equals the unconditional forecast-error covariance used in the notation of the question.
For completeness, derive the filtering step rather than leave the algorithm unspecified. Set and . Conditionally on the previous data, the observation has mean and covariance
Define the innovation process and gain by
The state forecast error and innovation are jointly normal, with cross-covariance . The residual is uncorrelated with by the definition of , hence independent of it by joint normality. Its mean is zero and its covariance is . This proves the conditional multivariate normal distribution update
Initialize . For , predict using the first boxed pair, store that forecast, form , and update using the second boxed pair. The updated state then starts the next iteration. This is the Kalman filter, and supplies the entire forecast sequence in one forward pass.
The requested unconditional innovation moments are
In fact , with deterministic for fixed parameters. Thus the innovation is independent of the past observations, and hence the innovations are mutually independent. Equivalently, orthogonality to the past together with joint normality gives this independence.
The chain rule for densities now yields the Gaussian innovation likelihood, treating the specified initial mean and covariance as known:
Both and are functions of the trial parameters, so the filtering recursion must be run again for each candidate . The ordinary density formula assumes positive definite; for example positive definite ensures this. Singular observation laws instead require densities on their Gaussian support.
Evaluate the log-likelihood numerically by the forward recursion, and maximize it over admissible , enforcing the covariance constraints. A Cholesky decomposition gives and the quadratic term , avoiding an explicit matrix inverse. Covariance parameterizations or constrained optimization keep positive semidefinite and the innovation covariance nonsingular. An attained maximizer is a maximum-likelihood estimator; in general this is a numerical optimization problem and neither uniqueness nor global convergence of a local optimizer is automatic.
Let be the total number of rejected null hypotheses, and let count rejected null hypotheses that are in fact true. The false discovery proportion and false discovery rate are
Thus the false discovery proportion is zero when there are no rejections, and the false discovery rate averages the realized proportion of erroneous rejections.
For the Benjamini-Hochberg procedure, order the p-values as and define
Reject every hypothesis whose p-value satisfies when , and reject none otherwise. There are exactly such hypotheses: if a further p-value also met the threshold, its larger rank would satisfy its own, still larger, threshold, contradicting maximality. Under the stated independence and uniform true-null p-values, this procedure has false discovery rate .
For the first modified Benjamini-Hochberg procedure, remove , order the remaining p-values as , and use the shifted critical values . Define its rejection count by
Equivalently, insert a zero in place of and run the ordinary Benjamini-Hochberg procedure on values; its total rejection count is . For , the Benjamini-Hochberg leave-one-out identity in event form is
To prove it, suppose first that the left event holds. Then is among the rejected values. Lowering it to zero leaves all ordered p-values at ranks greater than unchanged, so no new rank can meet its threshold. Rank still meets its threshold, giving . Conversely, if and , the rejected remaining values and all meet the original rank- threshold. Hence the original count is at least . Lowering a p-value cannot decrease the rejection count, so the original count is also at most . This proves equality of the events, including possible ties.
For the second modified Benjamini-Hochberg procedure, remove both , order the remaining p-values as , and use shifted critical values . Its count is
Replacing by zeros in the full procedure gives a total of rejections. The same argument gives the Benjamini-Hochberg leave-two-out identity, for :
In the forward direction both replaced values already lie among the first , so ordered ranks greater than remain unchanged. In the reverse direction reinserting the two values below the rank- threshold still leaves at least qualifying values, while monotonicity under lowering bounds the original count above by . For , the event of two p-values meeting the displayed threshold while only one is rejected is empty. For only the first modification is needed.
To calculate the second moment, write for a true null. Expanding the square gives
For a true-null index , let denote the first modified count after removing . It is a function only of the remaining p-values, and so is independent of the uniform variable . The first event identity therefore yields
For distinct true-null indices , the two independent uniform p-values are independent of the second modified count . The second event identity gives
The last sum is one because always takes a value in . The ordered pairs in the expansion of number , with no factor of one half.
The true-null p-values are identically distributed and independent, so permutations among their indices leave the law of the full collection unchanged. Thus every with has the same distribution as , even though the false-null p-values need not be identically distributed. Combining the diagonal and off-diagonal contributions proves the second moment of the Benjamini-Hochberg false discovery proportion:
If the first modified count is instead defined to include the inserted zero, the same answer is . When , both sides are zero; when , there are no off-diagonal terms. Finally, repeating the first calculation with denominator rather than its square gives , which verifies the claimed exact false discovery rate.
In a linear regression model, a large number of predictors or nearly dependent columns of the design matrix can make some eigenvalues of small. The ordinary least squares estimate has covariance matrix , so these directions have large variance, even though the estimator is unbiased. Ridge regression reduces this variance by shrinking the estimate towards zero, at the cost of introducing bias. This bias-variance tradeoff can improve the mean squared error when the signal is not too large in the shrunken directions.
For the usual preprocessing with a regression intercept, separate the constant column from the predictors and leave the intercept unpenalized. Subtract the sample mean from the response and from each nonconstant predictor, then divide each centered predictor column by its positive scale, for example its Euclidean norm. Thus all predictor columns have squared norm one; using squared norm instead simply changes the convention for the penalty. Centering is performed on the slope columns, not on the constant column. A full-rank design containing an intercept retains full rank on its centered slope columns. A model without an intercept should retain its specified mean structure unless an intercept is deliberately introduced.
Let be the resulting scaled, centered slope design matrix, let be the centered response, and let denote the coefficients in these standardized coordinates. The ridge regression problem is
Here is the number of slopes; it is one less than the original number of columns when a constant column was present. The gradient of this objective is . Its Hessian matrix is a positive-definite matrix, giving the unique closed-form ridge regression estimator
The regression intercept in the original coordinates is recovered as , and the original slopes are , where is the scale used for column . This is the unpenalized intercept in ridge regression convention. For a fixed full-rank coefficient design written simply as , with response and coefficient , the same formula reads
The risk comparison below first uses the fixed coefficient coordinates in which this quadratic penalty is imposed, and then addresses reversal of the scaling.
Assume at least one penalized coefficient, zero-mean errors with covariance matrix , and . No normal distribution assumption is needed for this mean squared error calculation. Put and . The covariance and bias of a ridge regression estimator follow directly from its linear expression:
For the centered slope model, the response errors have covariance matrix , where . Since , the same coefficient covariance matrix formula still holds with .
By the spectral theorem for real symmetric matrices, write , with orthogonal and every , and put . The bias-variance decomposition of mean squared error now gives
The expression at is the ordinary least squares mean squared error. Its right derivative is
Therefore, for every fixed coefficient vector, all sufficiently small positive penalties strictly improve mean squared error.
An explicit sufficient choice follows by subtracting the ordinary least squares risk term by term:
Since , every summand is negative whenever
When , every positive penalty works. This is a pointwise result: the permitted penalty may depend on the true coefficient vector. A fixed positive penalty cannot improve risk for all arbitrarily large signals, because its squared bias grows quadratically with the signal. The same distinction appears in uniform directional risk improvement by ridge regression and unbounded directional risk of fixed ridge shrinkage. In centered coordinates the unpenalized intercept contributes the same to both coefficient risks, so it does not affect the improvement. Reversing a fixed predictor scaling measures coefficient error with a fixed positive-definite matrix instead of the identity matrix. For this weighted loss the squared bias still has derivative zero at , while the variance term has derivative
The inequality follows because . Thus sufficiently small penalties also improve the original-coordinate mean squared error, though the explicit sufficient bound above was for the unweighted coordinates. Recovering an uncentered intercept adds a fixed nonnegative quadratic form to the slope loss, and its sample-mean error is uncorrelated with the centered slope errors, so the same weighted argument applies. A model with only an unpenalized intercept gives identical estimators. If the noise variance were zero, the strict improvement assertion would fail; positive noise variance is essential.
A real functional is Hadamard differentiable at if there exists a continuous linear functional such that, whenever , and in the normed vector space,
If is defined only on a subset, require to lie in its domain. The variation of the directions is essential: convergence along each fixed direction alone is a weaker differentiability condition. Tangential Hadamard differentiability restricts the limit directions to a specified subspace; here the derivative is defined on all of .
We prove the functional delta method using compact sets, so do not need an unjustified application of the Skorokhod representation theorem on an arbitrary nonseparable normed vector space. As usual for function-space convergence in distribution, the limiting random element is tight, and the real statistics are measurable, or convergence is understood through outer expectations. These are automatic under the usual separable complete random-element formulation; the empirical-process formulation used below has a tight separably supported limit.
For put
Hadamard differentiability says for every convergent sequence of directions, since continuity gives . It implies a Hadamard remainder near a compact set: for every compact and there are and such that
To prove it, suppose it failed. Choose and at distance less than from with . Choose with . A subsequence of converges by compactness, so the corresponding converge too. The defining derivative limit contradicts the lower bound on .
Now let and . Given , tightness of a probability measure provides a compact with . For the open -neighbourhood , the Portmanteau theorem gives
On the preceding uniform remainder is less than for all sufficiently large . Consequently . Since is arbitrary, this remainder tends to zero in probability. We have proved the actual expansion
Continuity of and the continuous mapping theorem give . The vanishing remainder and Slutsky theorem therefore yield
This derivation uses the derivative definition, rather than assuming a uniform Fréchet remainder over every bounded set. If outer expectations are needed, the same compact-neighbourhood argument uses outer probabilities and gives the corresponding empirical-process functional delta method.
For the statistical application, use the sample's empirical distribution function and the plug-in estimator
Here the printed denotes bounded pointwise functions with the supremum norm, namely , rather than equivalence classes modulo Lebesgue null sets. Assume Hadamard differentiability at the true and the usual measurability of the scalar estimator. The Donsker theorem for empirical distribution functions and the proved functional delta method give
No continuity of is needed for this application. Since the derivative is continuous and linear and the bridge is a centred Gaussian process, is a centred Gaussian real random variable, possibly identically zero.
Finally, a finite real weak limit ensures the required boundedness in probability. Choose so large that . Applying the closed-set part of the Portmanteau theorem to gives
Hence there is such that, for every ,
In particular the upper-tail condition in the question holds. The plug-in estimator has estimation error , even when its limiting variance is zero.
For an orthonormal wavelet construction, set
The scaling function generates the approximation spaces , while generates the detail spaces in the multiresolution analysis, with . At coarse resolution zero, the family is an orthonormal basis of .
For the given square-integrable function the coefficients and wavelet series are
For real wavelets these are the real L2 inner products; use complex conjugates for a complex basis. The series converges in the L2 norm, and its coefficients satisfy the Parseval identity for a Hilbertian basis
An hypothesis alone does not assert pointwise convergence.
The resolution- wavelet projection is the orthogonal projection onto :
The two expressions coincide by iterating . Orthogonality gives , which tends to zero. One can equivalently write the projection with kernel , interpreted as the same orthogonal projection.
In fixed-design nonparametric regression, the observations satisfy
with independent errors. In the usual Gaussian version . The regression function is evaluated at the design sites, so it must be specified pointwise there, not only as an equivalence class. For , the observations concern on . They cannot identify its values outside this interval. Use an interval-adapted wavelet basis on , consisting of coarse functions and detail functions ; a Haar wavelet basis is one explicit choice.
Here is a wavelet regression estimator valid for every . Divide into cells , with the value at zero assigned arbitrarily, and form the observed step function on . For any normalized basis function , estimate its coefficient by
The wavelet regression estimator at resolution is therefore
Only finitely many interval basis functions occur. Its expectation is , where equals on . Thus approximation of includes the resolution bias and the grid discretization bias. For continuous , the latter tends uniformly to zero by uniform continuity. No smoothness or consistency claim for arbitrary pointwise representatives of functions is needed to define the estimator.
The common point-evaluation version replaces each weight by , giving
These are Riemann-sum approximations to the coefficient integrals. For Haar wavelets, , and , they equal the integrated-weight coefficients exactly if the dyadic cells and pointwise basis versions are chosen right-closed. Each retained basis function is then constant on every sampling cell. This gives a concrete implementation by the discrete wavelet transform. A sensible finest resolution has ; beyond the grid scale there are details the data cannot resolve.
For wavelet coefficient thresholding, retain the coarse coefficients but discard detail coefficients dominated by noise. A hard-threshold version is
Alternatively, soft thresholding replaces each detail coefficient by . The unthresholded wavelet regression estimator retains all details at the chosen resolution; thresholding retains evidence for local features while suppressing small fluctuations.
The thresholds can be derived from a simultaneous wavelet coefficient noise bound. Write . Independence gives
where the inequality applies the Cauchy-Schwarz inequality on each cell of length . In the dyadic Haar implementation, exactly. For Gaussian errors, the normalized coefficient noise is standard normal. More generally, independent centred sub-Gaussian random variables with parameter give
If detail coefficients are considered, a union bound consequently justifies
With probability at least , every retained detail's noise is below its corresponding threshold; independence between the coefficient errors is unnecessary. With of order and , these thresholds have the familiar order . Coefficients larger than their noise level represent potential signal, and a coefficient zero in the discretized mean is not selected on this simultaneous event. If is unknown, a consistent noise-scale estimate can be substituted; for Gaussian errors and predominantly negligible finest-level signal, the median absolute finest-level coefficients divided by the standard-normal absolute median estimates . Without a Gaussian or sub-Gaussian tail assumption, finite variance alone does not justify the logarithmic thresholds: Chebyshev inequality and the union bound instead permit .

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact