Type: Package
Title: Estimation and Inference of Two-Way pAUC, pAUC and pODC
Version: 2.1.1
Date: 2017-1-12
Author: Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao
Maintainer: Xiang Lyu <lyu17@purdue.edu>
Description: Tools for estimating and inferring two-way partial area under receiver operating characteristic curves (two-way pAUC), partial area under receiver operating characteristic curves (pAUC), and partial area under ordinal dominance curves (pODC). Methods includes Mann-Whitney statistic and Jackknife, etc.
Imports: pROC, stats, graphics
Depends: R (≥ 3.1.1)
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
LazyData: true
RoxygenNote: 5.0.1
URL: http://arxiv.org/abs/1508.00298 http://www3.stat.sinica.edu.tw/statistica/j27n1/j27n117/j27n117.html http://www.ncbi.nlm.nih.gov/pubmed/20729218
Suggests: knitr, rmarkdown
VignetteBuilder: knitr
NeedsCompilation: no
Packaged: 2017-04-09 00:48:04 UTC; xianglyu
Repository: CRAN
Date/Publication: 2017-04-09 13:38:34 UTC

Estimation and Inference of Two-Way Partial AUC, FPR partial AUC and FNR partial ODC

Description

Tools of estimation and inference of two-way partial AUC, FPR partial AUC and FNR partial ODC. Methods are proposed in Yang et al., 2016 and Yang et al., 2017, including Mann-Whitney Statistic, jackknife method, etc.

Details

Package: tpAUC
Type: Package
Date 2017-04-08
License: GPL (>= 2)

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.
Maintainer: Xiang Lyu <lyu17@purdue.edu>

References

Wang Z, Chang Y. Marker selection via maximizing the partial area under the ROC curve of linear risk scores[J]. Biostatistics, 2011, 12(2): 369-385.
Yang H, Lu K, Lyu X, Hu F. Two-Way Partial AUC and Its Properties[J]. arXiv:1508.00298, 2016.
Yang H, Lu K, Zhao Y. A nonparametric approach for partial areas under ROC curves and ordinal dominance curves. Statistica Sinica, 2017, 27: 357-371.

Partial ODC Estimation and Inference

Description

Estimate and infer the area of region under ODC curve with pre-specific FNR constraint (FNR-pODC). See Yang et al., 2017 for details.

Usage

podc(response, predictor, threshold = 0.9, method = "MW", ci = TRUE,
  cp = 0.95, smooth = FALSE)

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0/1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

threshold

numeric; false negative rate (FNR) constraint.

method

methods to estimate FNR-pODC. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

ci

logic; compute the confidence interval of estimation?

cp

numeric; coverage probability of confidence interval.

smooth

if TRUE, the ODC curve is passed to smooth to be smoothed.

Details

This function estimates and infers FNR partial ODC given response, predictor and pre-specific FNR constraint. MW: Mann-Whitney statistic. expect: method in Yang et al., 2017 adapted from Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Value

Estimation and Inference of FNR partial ODC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

podc.est, podc.ci

Examples


library('pROC')
data(aSAH)
podc(aSAH$outcome, aSAH$s100b,threshold=0.9, method='expect',ci=TRUE, cp=0.95 )

Partial ODC Inference

Description

Infer the area of region under ordinal dominance curve with pre-specific FNR constraint (FNR-pODC). See Yang et al., 2017 for details.

Usage

podc.ci(response, predictor, cp = 0.95, threshold = 0.9, method = "MW")

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0 and 1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

cp

numeric; coverage probability of confidence interval.

threshold

numeric; false negative rate (FNR) constraint.

method

methods to estimate partial ODC. MW: Mann-Whitney statistic. expect: method in Yang et al., 2017 adapted from Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Details

This function infers FNR partial ODC given response, predictor and pre-specific FNR constraint. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Value

Confidence interval of FNR partial ODC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

proc.ci

Examples


library('pROC')
data(aSAH)
podc.ci(aSAH$outcome, aSAH$s100b, method='expect',threshold=0.8, cp=0.97)

Partial ODC Estimation

Description

Estimate the area of region under ordinal dominance curve with pre-specific FNR constraint (FNR-pODC). See Yang et al., 2017 for details.

Usage

podc.est(response, predictor, threshold = 0.9, method = "MW",
  smooth = FALSE)

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0 and 1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

threshold

numeric; false negative rate (FNR) constraint.

method

methods to estimate partial ODC. MW: Mann-Whitney statistic. expect: method in Yang et al., 2017 adapted from Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

smooth

if TRUE, the ODC curve is passed to smooth to be smoothed.

Details

This function estimates FNR partial ODC given response, predictor and pre-specific FNR constraint. MW: Mann-Whitney statistic. expect: method in Yang et al., 2017 adapted from Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Value

Estimation of FNR partial ODC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

proc.est

Examples


library('pROC')
data(aSAH)
podc.est(aSAH$outcome, aSAH$s100b, method='expect',threshold=0.8 )

Partial AUC Estimation and Inference

Description

Estimate and infer the area of region under ROC curve with pre-specific FPR constraint (FPR-pAUC). See Yang et al., 2017 for details.

Usage

proc(response, predictor, threshold = 0.9, method = "MW", ci = TRUE,
  cp = 0.95, smooth = FALSE)

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0/1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

threshold

numeric; false positive rate (FPR) constraint.

method

methods to estimate FPR-pAUC. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

ci

logic; compute the confidence interval of estimation?

cp

numeric; coverage probability of confidence interval.

smooth

if TRUE, the ROC curve is passed to smooth to be smoothed.

Details

This function estimates and infers FPR partial AUC given response, predictor and pre-specific FPR constraint. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Value

Estimate and Inference of FPR partial AUC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

roc, tproc.est, proc.est, proc.ci

Examples


library('pROC')
data(aSAH)
proc(aSAH$outcome, aSAH$s100b,threshold=0.9, method='expect',ci=TRUE, cp=0.95)

Partial AUC Inference

Description

Infer the area of region under ROC curve with pre-specific FPR constraint (FPR-pAUC). See Yang et al., 2017 for details.

Usage

proc.ci(response, predictor, cp = 0.95, threshold = 0.9, method = "MW")

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0/1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

cp

numeric; coverage probability of confidence interval.

threshold

numeric; false positive rate (FPR) constraint.

method

methods to estimate FPR-pAUC. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Details

This function infers FPR partial AUC given response, predictor and pre-specific FPR constraint. MW: Mann-Whitney statistic. method in Yang et al., 2017 adapted from Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Value

Confidence interval of FPR partial AUC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

tproc.est, podc.ci

Examples


library('pROC')
data(aSAH)
proc.ci(aSAH$outcome, aSAH$s100b, cp=0.95 ,threshold=0.9,method='expect')

Partial AUC Estimation

Description

Estimate the area of region under ROC curve with pre-specific FPR constraint (FPR-pAUC). See Yang et al., 2017 for details.

Usage

proc.est(response, predictor, threshold = 0.9, method = "MW",
  smooth = FALSE)

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0 and 1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

threshold

numeric; false positive rate (FPR) constraint.

method

methods to estimate FPR-pAUC. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

smooth

if TRUE, the ROC curve is passed to smooth to be smoothed.

Details

This function estimates FPR partial AUC given response, predictor and pre-specific FPR constraint. MW: Mann-Whitney statistic. expect: method in (2.2) Wang and Chang, 2011. jackknife: jackknife method in Yang et al., 2017.

Value

Estimate of FPR partial AUC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

tproc.est, podc.est

Examples


library('pROC')
data(aSAH)
proc.est(aSAH$outcome, aSAH$s100b, method='expect',threshold=0.8)

Two-Way Partial AUC Estimation

Description

Estimate the area of region under ROC curve under pre-specific FPR/TPR constraints (two-way partial AUC). See Yang et al., 2016 for details.

Usage

tproc.est(response, predictor, threshold = c(1, 0), smooth = FALSE)

Arguments

response

a factor, numeric or character vector of responses; typically encoded with 0 (negative) and 1 (positive). Only two classes can be used in a ROC curve. If its levels are not 0 and 1, the first level will be defaultly regarded as negative.

predictor

a numeric vector of the same length than response, containing the predicted value of each observation. An ordered factor is coerced to a numeric.

threshold

a length-two numeric vector; the first element is FPR threshold, the second is TPR.

smooth

if TRUE, the ROC curve is passed to smooth to be smoothed.

Details

This function estimates two-way partial AUC given response, predictor and pre-specific FPR/TPR constraints.

Value

Estimate of two-way partial AUC.

Author(s)

Hanfang Yang, Kun Lu, Xiang Lyu, Feifang Hu, Yichuan Zhao.

See Also

roc, podc.est, proc.est

Examples


library('pROC')
data(aSAH)
tproc.est(aSAH$outcome, aSAH$s100b, threshold=c(0.8,0.2))