cenroc 0.0.6

Calculation of censored ROC curve

censoredROC

This function computes the time-dependent ROC curve for right censored survival data using the cumulative sensitivity and dynamic specificity definitions. The ROC curves can be either empirical (non-smoothed) or smoothed with/wtihout boundary correction. It also calculates the time-dependent area under the ROC curve (AUC).

This package was built taken the cenROC package in R (https://cran.rstudio.com/web/packages/cenROC/index.html) as a main reference. The main idea was to allow Python users to apply the same tools.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

cenroc requires following packages to be installed

['numpy', 'pandas', 'matplotlib', 'scipy','statsmodels']

if you don't have them installed cenroc will install the latest versions of these packages on your machine

Installing

Use the following command to install the cenroc package from PyPi

pip install cenroc

Main class

The main class of this package is cenROC

In order to use the methods you have to intialise this class

cenROC(Y, M, censor, t, U = NULL, h = NULL, bw = "NR", method = "tra", ktype = "normal", ktype1 = "normal", B = 0, alpha = 0.05, plot = "TRUE")

Arguments

Y

The numeric vector of event-times or observed times.

M

The numeric vector of marker values for which the time-dependent ROC curves is computed.

censor

The censoring indicator, 1 if event, 0 otherwise.

t

A scaler time point at which the time-dependent ROC curve is computed.

U

The vector of grid points where the ROC curve is estimated. The default is a sequence of 151 numbers between 0 and 1.

h

A scaler for the bandwidth of Beran's weight calculaions. The defualt is the value obtained by using the method of Sheather and Jones (1991).

bw

A character string specifying the bandwidth estimation method for the ROC itself. The default is the "NR" normal reference method. The user can also introduce a numerical value.

method

The method of ROC curve estimation. The possible options are "emp" emperical metod; "untra" smooth without boundary correction and "tra" is smooth ROC curve estimation with boundary correction. The default is the "tra" smooth ROC curve estimate with boundary correction.

ktype

A character string giving the type kernel distribution to be used for smoothing the ROC curve: "normal", "epanechnikov", "biweight", or "triweight". By default, the "normal" kernel is used.

ktype1

A character string specifying the desired kernel needed for Beran weight calculation. The possible options are "normal", "epanechnikov", "tricube", "boxcar", "triangular", or "quartic". The defaults is "normal" kernel density. this one is not used currently

B

The number of bootstrap samples to be used for variance estimation. The default is 0, no variance estimation.

alpha

The significance level. The default is 0.05.

Methods

ROC()

Produces a numpy array with ROC estimations

AUC()

Produces a float showing AUC estimate

plot()

Plots the bootstrapped plot of ROC

Example

Install the lifelines package to import the dataset

pip install lifelines

Import the datasets from the lifelines package along with cenroc package

import lifelines.datasets as data
from cenroc import cenROC

df_test = data.load_panel_test()

cenROC_test1 = cenROC(Y=df_test['t'], M=df_test['var2'], censor=df_test['E'],
                     t=3, U=None, h=None, bw='NR', method="tra",
                     ktype="normal", ktype1="normal", B=3, alpha=0.05)

print(cenROC_test1.ROC())

Output

0.110795
0.180722
0.212862
0.235690
0.253942
0.269407
0.282974
0.295154
0.306269
0.316538
0.326117
...

cenROC_test2 = cenROC(Y=df_test['t'], M=df_test['var1'], censor=df_test['E'],
                                    t=2, U=None, h=None, bw=0.1, method="untra",
                                    ktype="epanechnikov", ktype1="normal", B=2, alpha=0.05)

print(cenROC_test2.AUC())

Output

0.6742424242424243

Due to the stochastic nature of bootstraping the graphs will be different each time. We can set seed to produce the same result for demonstartion purposes.

import numpy as np
np.random.seed(3)
cenROC_test1.plot()

Output

!

Contributing

Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.

Versioning

We use SemVer for versioning. For the versions available, see the tags on this repository.

Authors

This package was developed by plaindata.ai

• Yury Moskaltsov - Initial programming and package building - YuryMoskaltsov
• Miguel Pereira - Mathematical analysis, project oversight

See also the list of contributors who participated in this project.

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Acknowledgments

• Hat tip to anyone whose code was used
• Inspiration
• etc

Potential bugs and improvements

• Figure out translation of C function in Python correctly. Currently our function is calculating based on pure censored data instead of estimated conditional probabilities.
• Youden optimal cutpoint metric is not the same in R and Python, although they are very similar. This is due to the discrepancy in the interpolation functions. scipy.interpolate.interp1d() in Python and approx() in R. This should be tested more thoroughly to achieve identical results. All other metrics in Youden function are identical.