Description Details Note Author(s) References See Also Examples
This package offers three outlier detection algorithms for censored data using quantile regression.
Package: | OutlierDC |
Type: | Package |
Version: | 0.3-0 |
Date: | 2014-03-23 |
License: | GPL version 3 |
LazyLoad: | no |
We would like to thank Huxia Judy Wang and Lan Wang for permission to use their LCRQ functions.
Soo-Heang Eo and HyungJun Cho Maintainer: Soo-Heang Eo <eo.sooheang@gmail.com>
Eo S-H, Hong S-M Hong, Cho H (2014). Identification of outlying observations with quantile regression for censored data, Submitted.
Wang HJ, Wang L (2009) Locally Weighted Censored Quantile Regression. JASA 104:1117–1128. doi: 10.1198/jasa.2009.tm08230
odc
, plot
, coef
, show
, quantreg
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 | ## Not run:
require(OutlierDC)
# Toy example
data(ebd)
# The data consists of 402 observations with 6 variables.
dim(ebd)
# To show the first six observations of the dataset,
head(ebd)
#scoring algorithm
fit <- odc(Surv(log(time), status) ~ meta, data = ebd)
fit
coef(fit)
plot(fit)
# Add upper bound for the selection of outleirs
fit1 <- update(fit, k_s = 4)
fit1
plot(fit1)
# residual-based algorithm
fit2 <- odc(Surv(log(time), status) ~ meta, data = ebd, method = "residual", k_r = 1.5)
fit2
plot(fit2)
# To display all of outlying observations in the fitted object
fit2@outlier.data
# boxplot algorithm
fit3 <- odc(Surv(log(time), status) ~ meta, data = ebd, method = "boxplot", k_b = 1.5)
fit3
plot(fit3, ylab = "log survival times", xlab = "metastasis lymph nodes")
## End(Not run)
|
Loading required package: survival
Loading required package: quantreg
Loading required package: SparseM
Attaching package: 'SparseM'
The following object is masked from 'package:base':
backsolve
Attaching package: 'quantreg'
The following object is masked from 'package:survival':
untangle.specials
Loading required package: Formula
Package OutlierDC (0.3-0) loaded.
[1] 402 6
id meta exam status time ratio
1787 55468952 0 12 1 26 0.0000000
1788 8883016 0 12 1 11 0.0000000
1789 10647194 0 12 0 134 0.0000000
1790 16033679 2 12 1 1 0.1666667
1791 19519884 0 12 0 111 0.0000000
1792 19574077 0 12 1 8 0.0000000
Please wait...
Done.
Outlier Detection for Censored Data
Call: odc(formula = Surv(log(time), status) ~ meta, data = ebd)
Algorithm: Scoring algorithm (score)
Model: Locally weighted censored quantile regression (Wang)
Value for cut-off k_s:
# of outliers detected: 0
Top 6 outlying scores:
times delta (Intercept) meta score Outlier
346 4.48 0 1 9 4.59
327 2.71 1 1 13 4.54
326 2.08 1 1 14 2.52
296 4.86 1 1 4 2.35
354 3.09 1 1 10 2.11
233 5.29 0 1 1 1.95
q10 q25 q50 q75 q90
(Intercept) 1.632 2.565 3.401 4.500 5.196
meta -0.022 -0.077 -0.111 -0.183 -0.191
Outlier Detection for Censored Data
Call: odc(formula = Surv(log(time), status) ~ meta, data = ebd)
Algorithm: Scoring algorithm (score)
Model: Locally weighted censored quantile regression (Wang)
Value for cut-off k_s: 4
# of outliers detected: 2
Top 6 outlying scores:
times delta (Intercept) meta score Outlier
346 4.48 0 1 9 4.59 *
327 2.71 1 1 13 4.54 *
326 2.08 1 1 14 2.52
296 4.86 1 1 4 2.35
354 3.09 1 1 10 2.11
233 5.29 0 1 1 1.95
Please wait...
Done.
Outlier Detection for Censored Data
Call: odc(formula = Surv(log(time), status) ~ meta, data = ebd, method = "residual",
k_r = 1.5)
Algorithm: Residual-based algorithm (residual)
Model: Locally weighted censored quantile regression (Wang)
Value for cut-off k_r: 1.5
# of outliers detected: 9
Outliers detected:
times delta (Intercept) meta residual sigma Outlier
57 4.80 0 1 2 1.63 1.6 *
80 5.04 1 1 0 1.64 1.6 *
189 5.38 0 1 0 1.98 1.6 *
191 5.20 0 1 0 1.80 1.6 *
233 5.29 0 1 1 2.00 1.6 *
296 4.86 1 1 4 1.90 1.6 *
6 of all 9 outliers were displayed.
id meta exam status time ratio
57 39165334 2 13 0 122 0.15384615
80 2022934 0 13 1 154 0.00000000
189 25678892 0 16 0 217 0.00000000
191 10521031 0 17 0 181 0.00000000
233 52223267 1 18 0 198 0.05555556
296 27085350 4 20 1 129 0.20000000
346 12269804 9 24 0 88 0.37500000
357 17822095 0 25 1 157 0.00000000
395 43506173 0 37 0 152 0.00000000
Please wait...
Done.
Outlier Detection for Censored Data
Call: odc(formula = Surv(log(time), status) ~ meta, data = ebd, method = "boxplot",
k_b = 1.5)
Algorithm: Boxplot algorithm (boxplot)
Model: Locally weighted censored quantile regression (Wang)
Value for cut-off k_b: 1.5
# of outliers detected: 1
Outliers detected:
times delta (Intercept) meta UB Outlier
346 4.48 0 1 9 4.32 *
1 of all 1 outliers were displayed.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.