index.md

In topycyao/Xsurv: A package used to model survival data from advanced machine learning algorithms

Xsurv

Xsurv --- Efficient grdient boosting for survival data in R

The miracle of Xsurv

• Work with any type of survival data.
• One of the simplest syntax.
• Popular machine learning methods included (Xgboost, Lightgbm, random forests, and gbm).
• No parameters tuning bothering with cross-validation: everything can be done automatically! - Robust model determined risk levels.
• Know the important features associated with survival outcome:clearly presented by SHAP importance plot.

Let's Start

devtools::install_github("topycyao/Xsurv")          # Install the package

Table of Contents

• Introduction
• Contact
• Installation
• Features
• Quick Start
• Bugs and Issues
• Citation

Introduction

Xsurv is a useful and timely computational tool to identify candidate biomarkers that can predict or modulate patient prognosis, which may facilitate cancer translational and clinical research.

Contact

Create a ticket with a bug or question on GitHub Issues to help you and enrich it with your experience.

Installation

Latest release on GitHub

install.packages('devtools') # Ignore this if devtools is already installed.
devtools::install_github('topycyao/Xsurv')

Features

1. Easy manipulation of survival data:

   • The only thing you need to do is to divide data into two parts:covariates (x) and survival outcomes ;

   • Support different algorithms and keep updating;

   • No effort needed to tune your model even you have no experience:simply run Xsurv.cv and everything is done;

2. Prognostic biomarker discovery analysis made simple:

   • Directly know the top n features in fitted model;

   • A re-construct survival tree with important features helps to understand;

3. Model determined risk levels and robust predictions for survival probabilty:

Quick start

Load the package

library(Xsurv) #Load Xsurv into R

Xsurv can help to generate survival data

sim_dat<-Xsurv_sim_data(size=500,dim=20,lambda=2,vu=1, 
                                    c_rate=0.3)   # A data set is generated with sample size =500

#Covariates and survival outcome should be separted before fitting to Xsurv models
sim_x<-sim_dat[,1:20] # The first 20 (equal to dimension of covariates) columns
sim_y<-sim_dat[,c(21,22)] # The last 2 columns

Quick fit from data

fit<-Xsurv.cv(sim_x,sim_y,top_n=5)

Bugs and Issues

All bug reports, documentation improvements, enhancements and ideas are appreciated. Just let us know via GitHub.

Citation

Li,K.et al. (2021). Efficient gradient boosting for prognostic biomarker discovery.

topycyao/Xsurv documentation built on Aug. 6, 2022, 9:06 p.m.