pooledVarDataAccess

This function performs the data access step in the computation of a
differentially private pooled variance. The true values are computed using
the theoretical formula and <code><a href="/link/var?package=DPpack&version=0.2.0" data-mini-rdoc="DPpack::var">var</a></code>, while the sensitivities
are calculated based on bounded and unbounded differential privacy
Kifer2011DPpack according to the theoretical values
Liu2019bDPpack.

internal

An implementation of common statistical analysis and models with
differential privacy (Dwork et al., 2006a) <doi:10.1007/11681878_14>
guarantees. The package contains, for example, functions providing
differentially private computations of mean, variance, median, histograms,
and contingency tables. It also implements some statistical models and
machine learning algorithms such as linear regression (Kifer et al., 2012)
<https://proceedings.mlr.press/v23/kifer12.html>
and SVM (Chaudhuri et al., 2011)
<https://jmlr.org/papers/v12/chaudhuri11a.html>. In addition, it implements
some popular randomization mechanisms, including
the Laplace mechanism (Dwork et al., 2006a)
<doi:10.1007/11681878_14>, Gaussian mechanism (Dwork et al., 2006b)
<doi:10.1007/11761679_29>, analytic Gaussian mechanism (Balle & Wang, 2018)
<https://proceedings.mlr.press/v80/balle18a.html>, and exponential mechanism
(McSherry & Talwar, 2007) <doi:10.1109/FOCS.2007.66>.

Spencer Giddens

DPpack

Differentially Private Statistical Analysis and Machine Learning

Fang Liu

pooledVarDataAccess function

<dl><dt>samples</dt>
<dd>List of vectors from which to compute the pooled variance.</dd>
<dt>lower.bound</dt>
<dd>Real number giving the lower bound of the input data.</dd>
<dt>upper.bound</dt>
<dd>Real number giving the upper bound of the input data.</dd>
<dt>approx.n.max</dt>
<dd>Logical indicating whether to approximate n.max, which is
defined to be the length of the largest input vector. Approximation is best
if n.max is very large.</dd></dl>

Arguments

This function performs the data access step in the computation of a
differentially private pooled variance. The true values are computed using
the theoretical formula and <code><a href='https://rdrr.io/r/stats/cor.html'>var</a></code>, while the sensitivities
are calculated based on bounded and unbounded differential privacy
Kifer2011DPpack according to the theoretical values
Liu2019bDPpack.

Differentially Private Pooled Variance Data Access Function — pooledVarDataAccess

<dl>

<dt>samples</dt>
<dd>List of vectors from which to compute the pooled variance.</dd>


<dt>lower.bound</dt>
<dd>Real number giving the lower bound of the input data.</dd>


<dt>upper.bound</dt>
<dd>Real number giving the upper bound of the input data.</dd>


<dt>approx.n.max</dt>
<dd>Logical indicating whether to approximate n.max, which is
defined to be the length of the largest input vector. Approximation is best
if n.max is very large.</dd>

</dl>

pooledVarDataAccess: Differentially Private Pooled Variance Data Access Function

Description

Usage

Value

Arguments

References

Examples