• osx-64 v0.1.2.5

conda install

To install this package run one of the following:
conda install akode::survivalstan

---

survivalstan: Survival Models in Stan

author: Jacki Novik

Overview

Library of Stan Models for Survival Analysis

Features:

• Variety of standard survival models
  • Weibull, Exponential, and Gamma parameterization
  • PEM models with variety of baseline hazards
  • PEM model with varying-coefficients (by group)
  • PEM model with time-varying-effects
• Extensible framework - bring your own Stan code, or edit the models above
• Uses pandas data frames & patsy formulas
• Graphical posterior predictive checking (currently PEM models only)
• Plot posterior estimates of key parameters using seaborn
• Annotate posterior draws of parameter estimates, format as pandas dataframes
• Works with extensions to pystan, such as stancache or pystan-cache

Installation / Usage

Install using pip, as:

$ pip install survivalstan

Or, you can clone the repo:

$ git clone https://github.com/hammerlab/survivalstan.git
$ pip install .

Contributing

Details to come. For now, please do not hesitate to contribute if you would like.

Usage examples

There are several examples included in the example-notebooks, roughly one corresponding to each model.

If you are not sure where to start, Test pemsurvivalmodel with simulated data.ipynb contains the most explanatory text. Many of the other notebooks are sparse on explanation, but do illustrate variations on the different models.

For basic usage:

import survivalstan
import stanity
import seaborn as sb
import matplotlib.pyplot as plt
import statsmodels

## load flchain test data from R's `survival` package
dataset = statsmodels.datasets.get_rdataset(package = 'survival', dataname = 'flchain' )
d  = dataset.data.query('futime > 7')
d.reset_index(level = 0, inplace = True)

## e.g. fit Weibull survival model
testfit_wei = survivalstan.fit_stan_survival_model(
    model_cohort = 'Weibull model',
    model_code = survivalstan.models.weibull_survival_model,
    df = d,
    time_col = 'futime',
    event_col = 'death',
    formula = 'age + sex',
    iter = 3000,
    chains = 4,
    make_inits = survivalstan.make_weibull_survival_model_inits
    )

## coefplot for Weibull coefficient estimates
sb.boxplot(x = 'value', y = 'variable', data = testfit_wei['coefs'])

## or, use plot_coefs
survivalstan.utils.plot_coefs([testfit_wei])

## print summary of MCMC draws from posterior for each parameter
print(testfit_wei['fit'])


## e.g. fit Piecewise-exponential survival model 
dlong = survivalstan.prep_data_long_surv(d, time_col = 'futime', event_col = 'death')
testfit_pem = survivalstan.fit_stan_survival_model(
    model_cohort = 'PEM model',
    model_code = survivalstan.models.pem_survival_model,
    df = dlong,
    sample_col = 'index',
    timepoint_end_col = 'end_time',
    event_col = 'end_failure',
    formula = 'age + sex',
    iter = 3000,
    chains = 4,
    )

## print summary of MCMC draws from posterior for each parameter
print(testfit_pem['fit'])

## coefplot for PEM model results
sb.boxplot(x = 'value', y = 'variable', data = testfit_pem['coefs'])

## plot baseline hazard (only PEM models)
survivalstan.utils.plot_coefs([testfit_pem], element='baseline')

## posterior-predictive checking (only PEM models)
survivalstan.utils.plot_pp_survival([testfit_pem])

## e.g. compare models using PSIS-LOO
stanity.loo_compare(testfit_wei['loo'], testfit_pem['loo'])

## compare coefplots 
sb.boxplot(x = 'value', y = 'variable', hue = 'model_cohort',
    data = testfit_pem['coefs'].append(testfit_wei['coefs']))
plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)

## (or, use survivalstan.utils.plot_coefs)
survivalstan.utils.plot_coefs([testfit_wei, testfit_pem])