shared3p_statistics_summary.sc

shared3p_statistics_summary.sc

Module for finding the main percentiles in statistics.

Functions:

MAD

MAD(constant)

MAD(filter)

MAD(filter, constant)

covariance

covariance(filter)

fiveNumberSummary

maximum

mean

mean(filter)

minimum

quantiles

standardDev

standardDev(filter)

variance

variance(filter)

MAD

Find the median absolute deviation of a sample.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input sample

returns the median absolute deviation of the input multiplied by 1.4826

None

Function Overloads

D float32 MAD(D int32[[1]] data)

D float64 MAD(D int64[[1]] data)

D float32 MAD(D float32[[1]] data)

D float64 MAD(D float64[[1]] data)

MAD(constant)

Find the median absolute deviation of a sample.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input sample

constant

- scale factor

returns the median absolute deviation of the input multiplied by the scale factor

None

Function Overloads

D float32 MAD(D int32[[1]] data, float32 constant)

D float64 MAD(D int64[[1]] data, float64 constant)

D float32 MAD(D float32[[1]] data, float32 constant)

D float64 MAD(D float64[[1]] data, float64 constant)

MAD(filter)

Find the median absolute deviation of a filtered sample.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / int32 / int64

Parameters

data

- input sample

mask

- mask vector indicating which elements of the input sample to include when computing MAD

returns the median absolute deviation of the filtered input multiplied by 1.4826

Leaks the number of true values in the mask

Function Overloads

D float32 MAD(D int32[[1]] data, D bool[[1]] mask)

D float64 MAD(D int64[[1]] data, D bool[[1]] mask)

D float32 MAD(D float32[[1]] data, D bool[[1]] mask)

D float64 MAD(D float64[[1]] data, D bool[[1]] mask)

MAD(filter, constant)

Find the median absolute deviation of a filtered sample.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input sample

mask

- mask vector indicating which elements of the input sample to include when computing MAD

constant

- scale factor

returns the median absolute deviation of the filtered input multiplied by the scale factor

Leaks the number of true values in the mask

Function Overloads

D float32 MAD(D int32[[1]] data, D bool[[1]] mask, float32 constant)

D float64 MAD(D int64[[1]] data, D bool[[1]] mask, float64 constant)

D float32 MAD(D float32[[1]] data, D bool[[1]] mask, float32 constant)

D float64 MAD(D float64[[1]] data, D bool[[1]] mask, float64 constant)

covariance

Estimate the Pearson covariance of two samples.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

sample1

- first sample

sample2

- second sample are available

returns the covariance

None

Function Overloads

D float32 covariance(D int32[[1]] sample1, D int32[[1]] sample2)

D float64 covariance(D int64[[1]] sample1, D int64[[1]] sample2)

D float32 covariance(D float32[[1]] sample1, D float32[[1]] sample2)

D float64 covariance(D float64[[1]] sample1, D float64[[1]] sample2)

covariance(filter)

Estimate the Pearson covariance of two samples.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

sample1

- first sample

sample2

- second sample

filter

- filter indicating which elements of the samples are available

returns the covariance

None

Function Overloads

D float32 covariance(D int32[[1]] sample1, D int32[[1]] sample2, D bool[[1]] filter)

D float64 covariance(D int64[[1]] sample1, D int64[[1]] sample2, D bool[[1]] filter)

D float32 covariance(D float32[[1]] sample1, D float32[[1]] sample2, D bool[[1]] filter)

D float64 covariance(D float64[[1]] sample1, D float64[[1]] sample2, D bool[[1]] filter)

fiveNumberSummary

Find the minimum, lower quartile, median, upper quartile and maximum of a sample.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

A version of this function which hides the sample size was implemented for the paper "Going Beyond Millionaires: Privacy-Preserving Statistical Analysis" but is not included due to private integer division not working.

Parameters

data

- input sample

isAvailable

- vector indicating which elements of the input sample are available

returns a vector containing the minimum, lower quartile, median, upper quartile and maximum of the input sample (in that order). If the input size is less than five, a vector of zeroes is returned.

Leaks the number of true values in isAvailable

Function Overloads

D float32 fiveNumberSummary(D int32[[1]] data, D bool[[1]] isAvailable)

D float64 fiveNumberSummary(D int64[[1]] data, D bool[[1]] isAvailable)

D float32 fiveNumberSummary(D float32[[1]] data, D bool[[1]] isAvailable)

D float64 fiveNumberSummary(D float64[[1]] data, D bool[[1]] isAvailable)

maximum

Find the largest element of a vector.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input vector

isAvailable

- vector indicating which elements of the input vector are available

returns the largest element of the input vector

Leaks the number of true values in isAvailable

Function Overloads

D int32 maximum(D int32[[1]] data, D bool[[1]] isAvailable)

D int64 maximum(D int64[[1]] data, D bool[[1]] isAvailable)

D float32 maximum(D float32[[1]] data, D bool[[1]] isAvailable)

D float64 maximum(D float64[[1]] data, D bool[[1]] isAvailable)

mean

Find the mean of a vector.

Detailed Description

D - any protection domain

Supported types - int8 int16 int32 int64 uint8 uint16 uint32 uint64 float32 float64

Parameters

data

- input vector (the function may overflow if the input is too big)

returns the mean of the input vector

None

Function Overloads

D float32 mean(D int8[[1]] data)

D float32 mean(D int16[[1]] data)

D float32 mean(D int32[[1]] data)

D float64 mean(D int64[[1]] data)

D float32 mean(D uint8[[1]] data)

D float32 mean(D uint16[[1]] data)

D float32 mean(D uint32[[1]] data)

D float64 mean(D uint64[[1]] data)

D float32 mean(D float32[[1]] data)

D float64 mean(D float64[[1]] data)

mean(filter)

Find the mean of a filtered vector.

Detailed Description

D - any protection domain

Supported types - int8 int16 int32 int64 uint8 uint16 uint32 uint64 float32 float64

Parameters

data

- input vector (the function may overflow if the input is too big)

mask

- mask indicating which elements of the input vector to include when computing the mean

returns the mean of the filtered input vector

None

Function Overloads

D float32 mean(D int8[[1]] data, D bool[[1]] mask)

D float32 mean(D int16[[1]] data, D bool[[1]] mask)

D float32 mean(D int32[[1]] data, D bool[[1]] mask)

D float64 mean(D int64[[1]] data, D bool[[1]] mask)

D float32 mean(D uint8[[1]] data, D bool[[1]] mask)

D float32 mean(D uint16[[1]] data, D bool[[1]] mask)

D float32 mean(D uint32[[1]] data, D bool[[1]] mask)

D float64 mean(D uint64[[1]] data, D bool[[1]] mask)

D float32 mean(D float32[[1]] data, D bool[[1]] mask)

D float64 mean(D float64[[1]] data, D bool[[1]] mask)

minimum

Find the smallest element of a vector.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input vector

isAvailable

- vector indicating which elements of the input vector are available

returns the smallest element of the input vector

Leaks the number of true values in isAvailable

Function Overloads

D int32 minimum(D int32[[1]] data, D bool[[1]] isAvailable)

D int64 minimum(D int64[[1]] data, D bool[[1]] isAvailable)

D float32 minimum(D float32[[1]] data, D bool[[1]] isAvailable)

D float64 minimum(D float64[[1]] data, D bool[[1]] isAvailable)

quantiles

Calculate sample quantiles.

Detailed Description

D - shared3p protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input sample

isAvailable

- vector indicating which elements of the input sample are available

probabilities

- vector of probabilities between 0 and 1

returns sample quantiles corresponding to the specified probabilities

Leaks the number of true values in isAvailable

Function Overloads

D float32 quantiles(D int32[[1]] data, D bool[[1]] isAvailable, float64[[1]] probabilities)

D float64 quantiles(D int64[[1]] data, D bool[[1]] isAvailable, float64[[1]] probabilities)

D float32 quantiles(D float32[[1]] data, D bool[[1]] isAvailable, float64[[1]] probabilities)

D float64 quantiles(D float64[[1]] data, D bool[[1]] isAvailable, float64[[1]] probabilities)

standardDev

Find the standard deviation of a sample.

Detailed Description

D - any protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input sample (the function may overflow if the input is too big)

returns the standard deviation of the input sample

None

Function Overloads

D float32 standardDev(D int32[[1]] data)

D float64 standardDev(D int64[[1]] data)

D float32 standardDev(D float32[[1]] data)

D float64 standardDev(D float64[[1]] data)

standardDev(filter)

Find the standard deviation of a filtered sample.

Detailed Description

D - any protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input sample (the function may overflow if the input is too big)

mask

- mask vector indicating which elements of the input sample to include when computing the standard deviation

returns the standard deviation of the filtered input vector

None

Function Overloads

D float32 standardDev(D int32[[1]] data, D bool[[1]] mask)

D float64 standardDev(D int64[[1]] data, D bool[[1]] mask)

D float32 standardDev(D float32[[1]] data, D bool[[1]] mask)

D float64 standardDev(D float64[[1]] data, D bool[[1]] mask)

variance

Find the variance of a vector.

Detailed Description

D - any protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input vector (the function may overflow if the input is too big)

returns the variance of the input vector

None

Function Overloads

D float32 variance(D int32[[1]] data)

D float64 variance(D int64[[1]] data)

D float32 variance(D float32[[1]] data)

D float64 variance(D float64[[1]] data)

variance(filter)

Find the variance of a filtered vector.

Detailed Description

D - any protection domain

Supported types - int32 / int64 / float32 / float64

Parameters

data

- input vector (the function may overflow if the input is too big)

mask

- mask indicating which elements of the input vector to include when computing the variance

returns the variance of the input vector

None

Function Overloads

D float32 variance(D int32[[1]] data, D bool[[1]] mask)

D float64 variance(D int64[[1]] data, D bool[[1]] mask)

D float32 variance(D float32[[1]] data, D bool[[1]] mask)

D float64 variance(D float64[[1]] data, D bool[[1]] mask)