Thread Subject: need more object properties reference for bioinformatics toolbox

Do you have object reference for "classperf" function in the
bioinformatics toolbox? Properties that are now clear are:
Sampledistribution, Errordistribution, specificity,
sensitivity,etc..

"Arthur Zheng" <hzheng7@gatech.edu> wrote in message
<g36l8d$jgr$1@fred.mathworks.com>...
> Do you have object reference for "classperf" function in
the
> bioinformatics toolbox? Properties that are now clear are:
> Sampledistribution, Errordistribution, specificity,
> sensitivity,etc..

ValidationCounter: Counts the number of validations
performed.

SampleDistribution: A numeric vector that indicates how
many times every sample was considered into the validation.

ErrorDistribution: A numeric vector that indicates how many
times every sample was misclassified.

SampleDistributionByClass: A numeric vector that indicates
the frequency of the true classes in the validation.

ErrorDistributionByClass: A numeric vector that indicates
the frequency of errors for every class in the validation.

CountingMatrix: The classification confusion matrix. The
order of rows and columns is the same as GRP2IDX
(GROUNDTRUTH). Columns represent the true classes and rows
represent the classifier prediction. The last row in
CountingMatrix is reserved to count “inconclusive” results.
There are some families of classifiers that can reserve the
right to do make a hard class assignment, this can based on
metrics as the posterior probabilities or how close a
sample is to the class boundaries.

CorrectRate = CorrectlyClassifiedSamples /
ClassifiedSamples, note that inconclusive results are not
counted here.

ErrorRate = IncorrectlyClassifiedSamples /
ClassifiedSamples, note that inconclusive results are not
counted here.

InclonclusiveRate = NonClassifiedSamples /
TotalNumberOfSamples

ClassifiedRate = ClassifiedSamples / TotalNumberOfSamples

Sensitivity = CorrectlyClassifiedPositiveSamples /
TruePositiveSamples, inconclusive results which are true
positives are counted as errors for computing the
sensitivity (following a conservative approach), this means
considered as being incorrectly classified as negatives.

Specificity = CorrectlyClassifiedNegativeSamples /
TrueNegativeSamples, inconclusive results which are true
negatives are counted as errors for computing the
specificity (following a conservative approach), this means
considered as being incorrectly classified as positives.

PositivePredictiveValue =
CorrectlyClassifiedPositiveSamples /
PositiveClassifiedSamples, inconclusive results are
considered being classified as negatives for computing PPV

NegativePredictiveValue =
CorrectlyClassifiedNegativeSamples /
NegativeClassifiedSamples, inconclusive results are
considered being classified as positives for computing NPV

PositiveLikelihood = Sensitivity / (1- Specificity)

NegativeLikelihood = (1- Sensitivity) / Specificity

Prevalence = TruePositiveSamples / TotalNumberOfSamples