G***G
发帖数: 16778 | 1 I don't understand how to order the label in ROCR package in order to
get the correct result. I am wondering whether you can help me.
library(ROCR)
data(ROCR.simple)
pred <- prediction(ROCR.simple$predictions,ROCR.simple$labels)
svm.auc <- performance(pred, 'tpr', 'fpr')
plot(svm.auc)
#####label
ROCR.simple$labels[ROCR.simple$labels=="0"]="tumor"
ROCR.simple$labels[ROCR.simple$labels==1]="normal"
pred <- prediction(ROCR.simple$predictions,ROCR.simple$labels)
svm.auc <- performance(pred, 'tpr', ... 阅读全帖 |
|
G***G
发帖数: 16778 | 2 I don't understand how to order the label in ROCR package in order to
get the correct result. I am wondering whether you can help me.
library(ROCR)
data(ROCR.simple)
pred <- prediction(ROCR.simple$predictions,ROCR.simple$labels)
svm.auc <- performance(pred, 'tpr', 'fpr')
plot(svm.auc)
#####label
ROCR.simple$labels[ROCR.simple$labels=="0"]="tumor"
ROCR.simple$labels[ROCR.simple$labels==1]="normal"
pred <- prediction(ROCR.simple$predictions,ROCR.simple$labels)
svm.auc <- performance(pred, 'tpr', ... 阅读全帖 |
|
S******y
发帖数: 1123 | 3 I am using ROCR package in R for binary classification.
I am now using default to get ROC curves (true positive vs. false positive).
How can I set my own cutting points (i.e. if prediction prob > 0.65 then
classify as 1; else 0).
How does ROCR set its default cutting points?
Thanks! |
|
|
l**********1
发帖数: 5204 | 5 first BAOZI one got
Merci.
next,
rocr is better than roc
within obscure several
quantities that are necessary for evaluating the operational
effectiveness of diagnostic tests.
please go to
HTTPS//stat.ethz.ch/pipermail/r-help/2009-February/187606.html
and
//rocr.bioinf.mpi-sb.mpg.de/
now rocc might be better than rocr
//cran.r-project.org/web/packages/rocc/rocc.pdf
Ps: if you satisfied with this answer BAOZI another one please transfer to my account.
only for asking one old paper pdf file.
detai... 阅读全帖 |
|
g**********t
发帖数: 475 | 6 可以自己用bootstrap来计算某个点的CI。我没有用过ROCR来计算CI，不过我看文档
ROCR假定sensitivity/specificity服从正态分布（我不喜欢这个假定）。你也可以看
看qROC这个R library。里面有函数(ci.thresholds)来计算confidence interval。 |
|
a*********n
发帖数: 602 | 7 Washington DC revisiting in the end of year 2011
It has been 14 yrs since the first time I visited DC when I was an
International Relations’ student in Boston. I have lived and worked in DC
twice since, but have not been back in the past three years.
DC is always my most favorite city in US and I find this city getting more
attractive and charming. Streets are clearn, People are polite, friendly and
highly educated, cultural life is well orgnazied. And, I find out a lot of
buildings are popping ... 阅读全帖 |
|
G***G
发帖数: 16778 | 8 which one is better, roc or rocr? |
|
l**********1
发帖数: 5204 | 9 BAOZI one got
Merci.
回复]
[ 3 ]
发信人: GoooG (pumpkin), 信区: Biology
标 题: Re: roc
发信站: BBS 未名空间站 (Sat Feb 25 15:19:58 2012, 美东)
which one is better, roc or rocr? |
|
N******n
发帖数: 3003 | 10 tumor>normal
1>0,
你把本来的1，0 给翻了个因为tumor 当中1， normal 为0. |
|
G***G
发帖数: 16778 | 11 but why can we not label arbitrarily?
the prediton and label corresponds with each other very good.
evern our eyes can find the relationship. |
|
l**********1
发帖数: 5204 | 12 那需要你自己更新一下那个源程序喽 R
ps: it is not free lunch if you want other do that.
USD or funding share might needed. |
|
A*****n
发帖数: 243 | 13 install the ROCR package fom CRAN |
|
b*****a
发帖数: 905 | 14 lz自己挑著看吧。
## Make a ROC plot to see the ration of the rate of false positives to the
rate of false negatives
library(ROCR)
ip.glm<-glm(votesum ~ clint96 + partyid + aflcio97 + ccoal98, data=ip,
family=binomial(link=logit))
summary(ip.glm)
names(ip.glm)
pred.object<-prediction(ip.glm$fitted.values, ip.glm$y)
perf.object<-performance(pred.object, "tpr", "fpr")
# use add=TRUE as an argument to the plot() function if you want to overlay
additional ROC plots.
# Compare the model with submodel (Democr |
|
S******y
发帖数: 1123 | 15 Anybody know how to do a Gain Chart in R with these data points?
(I tried ROCR, it gave something like false/true positive, but it did not
give Gain Chart)
Thanks! |
|
D******n
发帖数: 2836 | 16 ROC was exactly invented to avoid setting up a cutoff.
Each point of the curve corresponds to one cutoff.
). |
|
|
z**********i
发帖数: 88 | 18 library(ROCR), YES. I don't know how to interprete the result.
~~~~~~~~~
Thanks, everybody |
|
D******n
发帖数: 2836 | 19 还是自己写个AUC的function吧，就几行。
我用那个rocr的居然会报错，不够内存，不知道那些人怎么写的。 |
|
i*****c
发帖数: 1322 | 20 I need to calculate the 95%CI of the sensitivity and specificity of a
cutpoint. I searched the internet, which said that I can use ROCR to get it.
Now I have the sensitivity and specificity for a serial cutpoint. Then how
to plot the ROC and get the statistics in R? Can anybody help me? Many
thanks. |
|
f*******n
发帖数: 2665 | 21 我刚开始学R，知道的class也就有dataframe, list之类的，下面的pred（pred <-
prediction()）是什么class类型？
> class(pred)
[1] "prediction"
attr(,"package")
[1] "ROCR" |
|
y*j
发帖数: 394 | 22 开车组
10am meet in front of
Rock Creek Nature Center and Planetarium (free parking)
5200 Glover Rood NW, Washington, DC 20015
(near the intersection of Military Road and Glover Road)
（兔子板斧能帮助组织carpool吗？）
10:05 出发走Rock Creek Park trail到动物园与Metro组回合, 晚到的就自己走赶上大
部队。
地图：http://www.nps.gov/pwr/customcf/apps/maps/showmap.cfm?
alphacode=rocr&parkname=Rock%20Creek%20Park
乘Metro组在动物园与开车组回合：
11:30am: National Zoo 熊猫馆 _出口外_大致不影响交通的地方（有更好的地方请建
议）有问题请联系john 703-622-1066
http://nationalzoo.si.edu/Visit/ZooMap/
... 阅读全帖 |
|
y*j
发帖数: 394 | 23 都是很容易的路, 逛公园,压马路. Nature Center -> Zoo 2-3miles (1-1.5 hours);
返回可以走不同的trails. Metro能到动物园.
天气预报最高气温47华氏度，注意保暖；地上会湿，请穿合适的鞋。
开车组
10am meet in front of
Rock Creek Nature Center and Planetarium (free parking)
5200 Glover Rood NW, Washington, DC 20015
(near the intersection of Military Road and Glover Road)
（兔子板斧能帮助组织carpool吗？）
10:05 出发走Rock Creek Park trail到动物园与Metro组回合, 晚到的就自己走赶上大
部队。
地图：http://www.nps.gov/pwr/customcf/apps/maps/showmap.cfm?
alphacode=rocr&parkname=Rock%20Creek%20Park
乘Metro组在动物园与开... 阅读全帖 |
|
