Cross-Sectional Idiosyncratic Diagnostics

Types

PortfolioOptimisers.IdioCalibrationSummaryResult — Type
struct IdioCalibrationSummaryResult{__T_mean_cs_std, __T_median_cs_std, __T_mean_kurtosis, __T_mean_skewness, __T_mean_tail_rate} <: AbstractResult

The five time-aggregated numbers of the calibration of a cross-sectional fit.

idio_calibration_summary returns it. Each number aggregates one of the calibration series over the observations that had enough assets to answer, and it is NaN when no observation did. A caller reads the fields by name, as s.mean_cs_std or (; mean_kurtosis) = s.

Fields

  • mean_cs_std: Mean of the calibration series of idio_calibration. Its reference under the normal law is 1.
  • median_cs_std: Median of the calibration series. Its reference under the normal law is 1.
  • mean_kurtosis: Mean of the excess kurtosis series of idio_kurtosis. Its reference under the normal law is 0.
  • mean_skewness: Mean of the skewness series of idio_skewness. Its reference under the normal law is 0.
  • mean_tail_rate: Mean of the tail rate series of idio_tail_rate. Its reference under the normal law is $2 \Phi(-c)$ at the threshold $c$.

Constructors

IdioCalibrationSummaryResult(mean_cs_std::Real, median_cs_std::Real,                             mean_kurtosis::Real, mean_skewness::Real,                             mean_tail_rate::Real)IdioCalibrationSummaryResult(; mean_cs_std::Real, median_cs_std::Real,                             mean_kurtosis::Real, mean_skewness::Real,                             mean_tail_rate::Real)

The positional constructor checks the values. The keyword constructor takes each field by name, with no default, and calls it.

Validation

  • mean_cs_std and median_cs_std are each NaN or not negative. Otherwise, the constructor raises a DomainError.
  • mean_tail_rate is NaN or in $[0, 1]$. Otherwise, the constructor raises a DomainError.

Related

source

Functions

PortfolioOptimisers.standardised_idio_returns — Function
standardised_idio_returns(eps::MatNum, vs::MatNum; ahead::Bool = true)
    -> Matrix{<:Real}

Return the standardised idiosyncratic returns of a cross-sectional fit.

Every calibration series of the idiosyncratic group reads it. The fit estimates a variance for each asset at each observation from the observations up to it, so the variance of observation $t - 1$ is the forecast for observation $t$. The standardised return states how large the realised return was against that forecast. A fit that is well calibrated leaves cross-sections of standardised returns whose standard deviation is 1.

The forecast must not have read the return it scores. The variance of observation $t$ has read $\varepsilon_{ti}$, so a division by it bounds the standardised return: an exponentially weighted variance with a weight $1 - \lambda$ on the latest squared return gives $|z_{ti}| \le 1 / \sqrt{1 - \lambda}$. That pulls the standard deviation below 1 and the tail rate below its Gaussian reference for a forecast that is right. So the default divides by the variance of the previous observation, and ahead = false gives the division by the variance of the same observation.

Mathematical definition

\[\begin{align} z_{ti} &= \begin{cases} \dfrac{\varepsilon_{ti}}{\hat{\sigma}_{t - 1, i}} & \text{if } \texttt{ahead}\,, \\ \dfrac{\varepsilon_{ti}}{\hat{\sigma}_{ti}} & \text{otherwise}\,. \end{cases} \end{align}\]

Where:

  • $z_{ti}$: Standardised idiosyncratic return of asset $i$ at observation $t$.
  • $\varepsilon_{ti}$: Idiosyncratic return of asset $i$ at observation $t$.
  • $\hat{\sigma}_{ti} = \sqrt{\max(v_{ti}, 0)}$: Predicted idiosyncratic volatility of asset $i$ at observation $t$. A negative variance counts as zero.
  • $v_{ti}$: Idiosyncratic variance of asset $i$ that the fit estimated from the observations up to $t$, so it is the forecast for observation $t + 1$.

An entry is NaN where the predicted volatility is not positive or the ratio is not finite, as standardised_idio_value states. Under ahead, the first observation has no forecast, so its row is NaN.

Arguments

  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • ahead: Whether each idiosyncratic return is divided by the volatility estimated before it. true divides the return of observation $t$ by the volatility of observation $t - 1$, whose estimate has not read that return, so a variance forecast that is right gives standardised returns of unit variance, and the first observation has no standardised return. false divides the return by the volatility of its own observation, whose estimate has read it, which pulls every standardised return toward zero and caps its size.

Validation

  • !isempty(eps).
  • size(vs) == size(eps).

Returns

  • z::Matrix{<:Real}: Standardised idiosyncratic returns observations × assets, of a floating-point type. An entry whose predicted volatility is zero, or whose return is not finite, is NaN, and under ahead the first row is NaN.

Examples

julia> standardised_idio_returns([1.0 -2.0; 3.0 4.0], [0.25 4.0; 1.0 0.0])2×2 Matrix{Float64}: NaN    NaN   6.0    2.0julia> standardised_idio_returns([1.0 -2.0; 3.0 4.0], [0.25 4.0; 1.0 0.0]; ahead = false)2×2 Matrix{Float64}: 2.0   -1.0 3.0  NaN

Related

source
PortfolioOptimisers.idio_calibration — Function
idio_calibration(z::MatNum) -> Vector{<:Real}
idio_calibration(eps::MatNum, vs::MatNum; ahead::Bool = true) -> Vector{<:Real}
idio_calibration(csfm::CrossSectionalFactorModel; ahead::Bool = true) -> Vector{<:Real}

Return the cross-sectional standard deviation of the standardised idiosyncratic returns, one entry per observation.

It is the headline calibration series of a cross-sectional fit. The fit predicted a variance for each asset. When the prediction is right, the realised idiosyncratic returns divided by the predicted volatilities have a standard deviation of 1. A series above 1 shows that the fit's specific risk is too small, and a series below 1 shows that it is too large.

Mathematical definition

\[\begin{align} \varsigma_{t} &= \sqrt{\frac{n_{t}}{n_{t} - 1} m_{2t}}\,. \end{align}\]

Where:

  • $\varsigma_{t}$: Sample standard deviation of the finite standardised returns of observation $t$.
  • $m_{pt}$: Central moment of order $p$ of the finite standardised returns of observation $t$.
  • $n_{t}$: Count of the assets in $\mathcal{F}_{t}$.
  • $\mathcal{F}_{t}$: Finite cross-section of observation $t$, the assets whose standardised return $z_{ti}$ is finite.
  • $z_{ti}$: Standardised idiosyncratic return of asset $i$ at observation $t$.

The deviation is the sample one, so it divides by $n_{t} - 1$. It is not defined for $n_{t} < 2$. A constant cross-section gives $\varsigma_{t} = 0$.

Arguments

  • z: Standardised idiosyncratic returns observations × assets.
  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • ahead: Whether each idiosyncratic return is divided by the volatility estimated before it. true divides the return of observation $t$ by the volatility of observation $t - 1$, whose estimate has not read that return, so a variance forecast that is right gives standardised returns of unit variance, and the first observation has no standardised return. false divides the return by the volatility of its own observation, whose estimate has read it, which pulls every standardised return toward zero and caps its size.

Validation

  • !isempty(eps) and size(vs) == size(eps), on the two-argument form.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.

Returns

  • c::Vector{<:Real}: The series, one entry per observation. An observation with fewer than two finite assets is NaN.

Related

source
PortfolioOptimisers.idio_tail_rate — Function
idio_tail_rate(z::MatNum; threshold::Real = 3) -> Vector{<:Real}
idio_tail_rate(eps::MatNum, vs::MatNum; threshold::Real = 3, ahead::Bool = true)
    -> Vector{<:Real}
idio_tail_rate(csfm::CrossSectionalFactorModel; threshold::Real = 3,
               ahead::Bool = true) -> Vector{<:Real}

Return the share of assets whose standardised idiosyncratic return exceeds a threshold, one entry per observation.

Under the normal law the expected rate is $2 \Phi(-c)$, which is about 0.0027 for a threshold of three. A series above it shows standardised returns with heavier tails than the normal law gives. The standardised returns of an equity factor model are often fat-tailed, so a rate above the Gaussian reference is common and is not by itself a defect of the fit.

Mathematical definition

\[\begin{align} r_{t} &= \frac{1}{n_{t}} \sum_{i \in \mathcal{F}_{t}} \mathbb{1} \left\{ \left| z_{ti} \right| > c \right\}\,. \end{align}\]

Where:

  • $r_{t}$: Tail rate of observation $t$.
  • $c$: The threshold.
  • $\Phi$: Cumulative distribution function of the standard normal law.
  • $z_{ti}$: Standardised idiosyncratic return of asset $i$ at observation $t$.
  • $\mathcal{F}_{t}$: Finite cross-section of observation $t$, the assets whose standardised return $z_{ti}$ is finite.
  • $n_{t}$: Count of the assets in $\mathcal{F}_{t}$.

The rate is not defined for $n_{t} = 0$. An entry that is not finite enters neither the count nor the sum.

Arguments

  • z: Standardised idiosyncratic returns observations × assets.
  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • threshold: Absolute standardised return above which an asset enters the rate.
  • ahead: Whether each idiosyncratic return is divided by the volatility estimated before it. true divides the return of observation $t$ by the volatility of observation $t - 1$, whose estimate has not read that return, so a variance forecast that is right gives standardised returns of unit variance, and the first observation has no standardised return. false divides the return by the volatility of its own observation, whose estimate has read it, which pulls every standardised return toward zero and caps its size.

Validation

  • !isempty(eps) and size(vs) == size(eps), on the two-argument form.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.

Returns

  • r::Vector{<:Real}: The series, one entry per observation. An observation with no finite asset is NaN.

Related

source
PortfolioOptimisers.idio_kurtosis — Function
idio_kurtosis(z::MatNum) -> Vector{<:Real}
idio_kurtosis(eps::MatNum, vs::MatNum; ahead::Bool = true) -> Vector{<:Real}
idio_kurtosis(csfm::CrossSectionalFactorModel; ahead::Bool = true) -> Vector{<:Real}

Return the cross-sectional excess kurtosis of the standardised idiosyncratic returns, one entry per observation.

The Gaussian reference is 0, and a positive value shows a cross-section with heavier tails than the normal law gives. Read it beside idio_tail_rate. The rate counts the assets beyond a threshold, and the kurtosis measures how far beyond it they went.

Mathematical definition

\[\begin{align} k_{t} &= \frac{n_{t} - 1}{(n_{t} - 2)(n_{t} - 3)} \left[ (n_{t} + 1) \left( \frac{m_{4t}}{m_{2t}^{2}} - 3 \right) + 6 \right]\,. \end{align}\]

Where:

  • $k_{t}$: Excess kurtosis of the finite standardised returns of observation $t$.
  • $m_{pt}$: Central moment of order $p$ of the finite standardised returns of observation $t$.
  • $n_{t}$: Count of the assets in $\mathcal{F}_{t}$.
  • $\mathcal{F}_{t}$: Finite cross-section of observation $t$, the assets whose standardised return $z_{ti}$ is finite.
  • $z_{ti}$: Standardised idiosyncratic return of asset $i$ at observation $t$.

It is the bias-corrected sample excess kurtosis, the estimator $G_{2}$ of [86]. It equals $k_{4} / k_{2}^{2}$, the fourth k-statistic over the square of the second, and the sample moment $m_{4t} / m_{2t}^{2} - 3$ is the estimator $g_{2}$ of the same work. It is not defined for $n_{t} < 4$, or for a constant cross-section, where $m_{2t} = 0$.

Arguments

  • z: Standardised idiosyncratic returns observations × assets.
  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • ahead: Whether each idiosyncratic return is divided by the volatility estimated before it. true divides the return of observation $t$ by the volatility of observation $t - 1$, whose estimate has not read that return, so a variance forecast that is right gives standardised returns of unit variance, and the first observation has no standardised return. false divides the return by the volatility of its own observation, whose estimate has read it, which pulls every standardised return toward zero and caps its size.

Validation

  • !isempty(eps) and size(vs) == size(eps), on the two-argument form.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.

Returns

  • k::Vector{<:Real}: The series, one entry per observation. An observation with fewer than four finite assets, or with a constant cross-section, is NaN.

Related

References

  • [86] D. N. Joanes and C. A. Gill. Comparing measures of sample skewness and kurtosis. Journal of the Royal Statistical Society: Series D (The Statistician) 47, 183–189 (1998). Pages 183 to 189. The original source of this formulation was not found. The formulation follows this work and is not checked against the original.
source
PortfolioOptimisers.idio_skewness — Function
idio_skewness(z::MatNum) -> Vector{<:Real}
idio_skewness(eps::MatNum, vs::MatNum; ahead::Bool = true) -> Vector{<:Real}
idio_skewness(csfm::CrossSectionalFactorModel; ahead::Bool = true) -> Vector{<:Real}

Return the cross-sectional skewness of the standardised idiosyncratic returns, one entry per observation.

The Gaussian reference is 0. A cross-section that stays skewed shows a residual with a direction the factors did not take. A sector or a style that the exposures do not name can cause it.

Mathematical definition

\[\begin{align} s_{t} &= \frac{m_{3t}}{m_{2t}^{3/2}} \frac{\sqrt{n_{t}(n_{t} - 1)}}{n_{t} - 2}\,. \end{align}\]

Where:

  • $s_{t}$: Skewness of the finite standardised returns of observation $t$.
  • $m_{pt}$: Central moment of order $p$ of the finite standardised returns of observation $t$.
  • $n_{t}$: Count of the assets in $\mathcal{F}_{t}$.
  • $\mathcal{F}_{t}$: Finite cross-section of observation $t$, the assets whose standardised return $z_{ti}$ is finite.
  • $z_{ti}$: Standardised idiosyncratic return of asset $i$ at observation $t$.

It is the bias-corrected sample skewness, the estimator $G_{1}$ of [86]. It equals $k_{3} / k_{2}^{3/2}$, the third k-statistic over the second to the power $3/2$, and the sample moment ratio $m_{3t} / m_{2t}^{3/2}$ is the estimator $g_{1}$ of the same work. It is not defined for $n_{t} < 3$, or for a constant cross-section, where $m_{2t} = 0$.

Arguments

  • z: Standardised idiosyncratic returns observations × assets.
  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • ahead: Whether each idiosyncratic return is divided by the volatility estimated before it. true divides the return of observation $t$ by the volatility of observation $t - 1$, whose estimate has not read that return, so a variance forecast that is right gives standardised returns of unit variance, and the first observation has no standardised return. false divides the return by the volatility of its own observation, whose estimate has read it, which pulls every standardised return toward zero and caps its size.

Validation

  • !isempty(eps) and size(vs) == size(eps), on the two-argument form.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.

Returns

  • s::Vector{<:Real}: The series, one entry per observation. An observation with fewer than three finite assets, or with a constant cross-section, is NaN.

Related

References

  • [86] D. N. Joanes and C. A. Gill. Comparing measures of sample skewness and kurtosis. Journal of the Royal Statistical Society: Series D (The Statistician) 47, 183–189 (1998). Pages 183 to 189. The original source of this formulation was not found. The formulation follows this work and is not checked against the original.
source
PortfolioOptimisers.idio_vol_ic — Function
idio_vol_ic(eps::MatNum, vs::MatNum; ties::Symbol = :average) -> Vector{<:Real}
idio_vol_ic(csfm::CrossSectionalFactorModel; ties::Symbol = :average) -> Vector{<:Real}

Return the information coefficient of the predicted idiosyncratic volatility, one entry per pair of observations.

The fit predicted a volatility for each asset at observation $t$. The assets it called the most volatile should be the assets that moved the most at observation $t + 1$. This series scores that ranking.

A high value shows a fit that ranks specific risk across the assets well. The series also responds to a broad cross-sectional effect such as size or liquidity. So read it beside idio_vol_residual_dependence, which shows whether the level of the prediction still leaks into the standardised returns.

Mathematical definition

\[\begin{align} \mathrm{IC}_{t} &= \rho_{S} \left( \hat{\sigma}_{t \cdot}, \left| \varepsilon_{t + 1, \cdot} \right| \right)\,. \end{align}\]

Where:

  • $\mathrm{IC}_{t}$: Information coefficient of observation $t$, read against observation $t + 1$.
  • $\hat{\sigma}_{t \cdot}$, $\varepsilon_{t + 1, \cdot}$: The cross-sections of observations $t$ and $t + 1$, one entry per asset.
  • $\rho_{S}$: Cross-sectional rank correlation, over the assets at which both cross-sections are finite. Under the default ties = :average, equal values share the mean of their ranks. Under ties = :ordinal, they take consecutive ranks, in the order of the asset axis.
  • $\hat{\sigma}_{ti} = \sqrt{\max(v_{ti}, 0)}$: Predicted idiosyncratic volatility of asset $i$ at observation $t$. A negative variance counts as zero.
  • $v_{ti}$: Idiosyncratic variance of asset $i$ that the fit estimated from the observations up to $t$, so it is the forecast for observation $t + 1$.
  • $\varepsilon_{ti}$: Idiosyncratic return of asset $i$ at observation $t$.

The correlation is not defined over fewer than five common finite assets.

Arguments

  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • ties: Rule that ranks the equal values of a cross-section. :average gives each value of a tie the mean of the positions the tie takes, so a rank correlation does not read the order of the assets, and a constant cross-section has no rank correlation. :ordinal gives the values of a tie consecutive positions, in the order of the asset axis. The two rules give the same ranks to a cross-section with no tie.

Validation

  • !isempty(eps), size(vs) == size(eps) and size(eps, 1) > 1.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.
  • The rules of cs_ranks.

Returns

  • c::Vector{<:Real}: The series, observations - 1 entries. An observation that shares fewer than five finite assets is NaN.

Related

source
PortfolioOptimisers.idio_vol_residual_dependence — Function
idio_vol_residual_dependence(eps::MatNum, vs::MatNum;
                             ties::Symbol = :average) -> Vector{<:Real}
idio_vol_residual_dependence(csfm::CrossSectionalFactorModel;
                             ties::Symbol = :average) -> Vector{<:Real}

Return the rank correlation of the predicted idiosyncratic volatility against the next observation's standardised absolute idiosyncratic return, one entry per pair of observations.

Division of the realised move by the predicted volatility should remove the level of the prediction. So a fit that is well calibrated leaves a series near 0, because the size of an asset's standardised move does not depend on how volatile the fit said the asset would be. A series that stays positive shows a fit that under-predicts the volatile assets, and a series that stays negative shows a fit that over-predicts them.

The target divides the absolute return of observation $t + 1$ by the volatility estimated at observation $t$, whose estimate has not read that return. It is the absolute standardised return $|z_{t + 1, i}|$ of standardised_idio_returns under its default ahead = true. The volatility of observation $t + 1$ has read the return, so a division by it would bound the target and pull it toward zero.

Read it beside idio_vol_ic. A fit that ranks well and leaves no residual dependence has a high information coefficient and a dependence near 0.

Mathematical definition

\[\begin{align} d_{t} &= \rho_{S} \left( \hat{\sigma}_{t \cdot}, \frac{\left| \varepsilon_{t + 1, \cdot} \right|}{\hat{\sigma}_{t \cdot}} \right)\,. \end{align}\]

Where:

  • $d_{t}$: Residual dependence of observation $t$, read against observation $t + 1$.
  • $\hat{\sigma}_{t \cdot}$, $\varepsilon_{t + 1, \cdot}$: The cross-sections of observations $t$ and $t + 1$, one entry per asset. The division is per asset, and it is NaN where $\hat{\sigma}_{ti}$ is zero.
  • $\rho_{S}$: Cross-sectional rank correlation, over the assets at which both cross-sections are finite. Under the default ties = :average, equal values share the mean of their ranks. Under ties = :ordinal, they take consecutive ranks, in the order of the asset axis.
  • $\hat{\sigma}_{ti} = \sqrt{\max(v_{ti}, 0)}$: Predicted idiosyncratic volatility of asset $i$ at observation $t$. A negative variance counts as zero.
  • $v_{ti}$: Idiosyncratic variance of asset $i$ that the fit estimated from the observations up to $t$, so it is the forecast for observation $t + 1$.
  • $\varepsilon_{ti}$: Idiosyncratic return of asset $i$ at observation $t$.

The correlation is not defined over fewer than five common finite assets.

Arguments

  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • ties: Rule that ranks the equal values of a cross-section. :average gives each value of a tie the mean of the positions the tie takes, so a rank correlation does not read the order of the assets, and a constant cross-section has no rank correlation. :ordinal gives the values of a tie consecutive positions, in the order of the asset axis. The two rules give the same ranks to a cross-section with no tie.

Validation

  • !isempty(eps), size(vs) == size(eps) and size(eps, 1) > 1.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.
  • The rules of cs_ranks.

Returns

  • d::Vector{<:Real}: The series, observations - 1 entries. An observation that shares fewer than five finite assets is NaN.

Related

source
PortfolioOptimisers.idio_calibration_summary — Function
idio_calibration_summary(eps::MatNum, vs::MatNum; threshold::Real = 3,
                         ahead::Bool = true) -> IdioCalibrationSummaryResult
idio_calibration_summary(csfm::CrossSectionalFactorModel; threshold::Real = 3,
                         ahead::Bool = true) -> IdioCalibrationSummaryResult

Return the five time-aggregated numbers of the calibration of a cross-sectional fit.

The four calibration series each answer per observation. To judge a whole fit, read their time aggregates. The five numbers are the mean and the median of idio_calibration, and the means of idio_kurtosis, idio_skewness and idio_tail_rate. Each aggregate skips the observations that had too few assets to answer.

Under the normal law the expected values are 1, 1, 0, 0 and $2 \Phi(-c)$, where $\Phi$ is the cumulative distribution function of the standard normal law and $c$ is the threshold. A fit of an equity universe usually has a positive excess kurtosis and a tail rate above the Gaussian reference. Read the first two numbers for the scale of the specific risk, and the last three for the shape of its tails.

Algorithm

  1. Compute the standardised returns z once with standardised_idio_returns.
  2. Compute the calibration series cs from z with idio_calibration.
  3. Take the mean of cs with idio_nan_mean and its median with idio_nan_median, giving mean_cs_std and median_cs_std.
  4. Compute the kurtosis, skewness and tail rate series from z, and take the mean of each with idio_nan_mean, giving mean_kurtosis, mean_skewness and mean_tail_rate.

Arguments

  • eps: Idiosyncratic return history observations × assets.
  • vs: Idiosyncratic variance history observations × assets.
  • csfm: A cross-sectional factor model block.
  • threshold: Absolute standardised return above which an asset enters the tail rate.
  • ahead: Whether each idiosyncratic return is divided by the volatility estimated before it. true divides the return of observation $t$ by the volatility of observation $t - 1$, whose estimate has not read that return, so a variance forecast that is right gives standardised returns of unit variance, and the first observation has no standardised return. false divides the return by the volatility of its own observation, whose estimate has read it, which pulls every standardised return toward zero and caps its size.

Validation

  • !isempty(eps) and size(vs) == size(eps), on the two-argument form.
  • csfm.csr is not nothing, else the verb raises an IsNothingError that names csr.
  • csfm.vs is not nothing, else the verb raises an IsNothingError that names vs.

Returns

  • s::IdioCalibrationSummaryResult: The five numbers below. Earlier releases returned a NamedTuple of the same fields.

    • mean_cs_std::Real: Mean of the calibration series.
    • median_cs_std::Real: Median of the calibration series.
    • mean_kurtosis::Real: Mean of the excess kurtosis series.
    • mean_skewness::Real: Mean of the skewness series.
    • mean_tail_rate::Real: Mean of the tail rate series.

Related

source

References

[86]
D. N. Joanes and C. A. Gill. Comparing measures of sample skewness and kurtosis. Journal of the Royal Statistical Society: Series D (The Statistician) 47, 183–189 (1998).