LADObj-class: Class '"LADObj"'

In Blossom: Statistical Comparisons with Distance-Function Based Permutation Tests

Description

The LADObj class holds model specification and output from the Least Absolute Deviation lad and Hypothesis Test hypothesis.test commands. The summary.LADObj class represents the summary of these objects.

Objects from the Class

Objects can be created by calls of the form new("LADObj", ...) or, more commonly, via the lad or hypothesis.test function call.

Slots

inputData:

A data.frame used for the analysis. The first column is the group (if supplied) and the second is the variable for which the analysis was performed.

NumObs:

The total number of observations in the original data set.

NumVars:

The total number of variables in the original data set.

theta:

The quantile used if quantile regression is desired.

NumPerm:

The number of permutations used.

Test:

A logical indicator of whether a hypothesis test was performed.

DoublePermutation:

A logical indicator of whether a double permutation was used this should only be TRUE for hypothesis.test.

T_o:

The dispersion test statistic equal to (sum of absolute deviations for reduced model - sum of absolute deviations for full model) / sum of absolute deviations for full model.

AsyRankScore:

If the option rank.score is set equal to true in the hypothesis test then this value contains the asymptotic rank score statistic.

P_value:

The probability of a smaller or equal delta under the null hypothesis.

P_valueTN:

The P-Value of Asymptotic rank score statistic.

Betas:

A vector of regression coefficients in the full model fit.

RedBetas:

A vector of regression coefficients in the reduced model fit (if applicable).

SumAbsValRes:

The sum of absolute values of the residuals in the full model.

SumAbsValResRed:

The sum of absolute values of the residuals in the reduced model (if applicable).

WtSumAbsDevsFulMod:

The weighted sum of the absolute deviations in the full model.

WtSumAbsDevsRedMod:

The weighted sum of the absolute deviations in the full model (if applicable).

NumIter:

The number of iterations for model convergence.

ExitCode:

A numeric code indicating whether a solution was successfully found:

• ExitCode = 1: Successful

• ExitCode = 2: Rounding Error

• ExitCode = 0: Non-Unique

• ExitCode = 7: Multiple Solutions

PermVals:

if save.test was set equal to TRUE and a hypothesis test was performed then this is a vector of the permutation values of the test statistic.

HasIntercept:

A logical vector indicating whether the full and reduced models include intercept terms.

DoAllQuants:

A logical indicator of whether all quantile regression estimates were calculated.

DoRankScore:

A logical indicator of whether a rank score test was performed.

IsOLS:

A logical indicator of whether ordinary least squares regression was performed.

NumLaVars:

A vector with the number of variables in the full and reduced models.

ResRed:

Residuals from the reduced model (if applicable).

Resids:

Residuals from the full model.

Call:

The original function call.

response:

The response vector used in the model fit.

full.mod.names:

The names of predictors used in the full model for hypothesis testing.

QuantOut:

If all.quants was true in the function code this is a matrix containing the quantile values. Using the function QuantValues() will display these in a useful manner.

Methods

coefficients

signature(x = "LADObj"): Returns the coefficients from the model fit.

predict

signature(x = "LADObj"): Reports the predicted values from the model fit.

print

signature(x = "LADObj"): Same as summary.

pvalue

signature(x = "LADObj"): Returns the p-value from the LadObj assuming a hypothesis test was specified.

QuantValues

signature(x = "LadObj"): Reports all quantile regression estimates if all.quants was set to TRUE.

ResampVals

signature(x = "LADObj"): Returns the resample values if applicable.

residuals

signature(x = "LADObj"): Returns the residuals from the full model fit.

show

signature(object = "LADObj"): Same as summary.

summary

signature(object = "LADObj"): Prints a detailed summary of the LAD fit.

Author(s)

Marian Talbert

References

Cade, B.S. and J.D. Richards. 1996. Permutation tests for least absolute deviation regression. Biometrics 52, 886–902.

See Also

lad, hypothesis.test

Examples

Out <- lad(lnlctm~widrat,data = lahontan,all.quants = TRUE)
summary(Out)
AllQuants <- QuantValues(Out)

Blossom documentation built on May 29, 2017, 10:55 p.m.