In AxelitoMartin/OncoCast: Ensemble learning survival prediction with improved interface.

knitr::opts_chunk$set(echo = F,fig.height = 8, fig.width = 12,message = F,warning = F)
library(knitr)

Summary

In this file we summarise the results from the oncocast run with the inputted data that had r params$n observations for r params$n.events events with r params$n.features features. You selected OncoCast using r params$method with r params$n.CV cross-validation.

OncoCast run

Prognostic performance

In this first section we go over the results shown in the "Dashboard" tab of the app. This summarises the average predicted risk distribution of patients, and the resulting prognostic power of the model

params$riskHist
kable(params$riskSum, row.names = T)
kable(params$CPE, row.names = T)
temp <- params$riskTable
colnames(temp) <- c("Coefficient","Hazard Ratio","S.E.","Z-value","P-value")
temp[1,] <- round(temp[1,],digits=4)
if(temp[,5] < 0.00001) temp[,5] <- "<0.00001"
kable(temp, row.names = T)

Risk group stratification

We show here the resulting risk groups from the following dichotomization: r paste0("RiskGroup",1:(length(params$cuts)+1),": ",c(0,params$cuts),"% to ",c(params$cuts,100),"% ",collapse = ",") of the averaged predicted risk score of the cohort. We present the Kaplan-Meier stratification associated with these risk groups:

params$KM
kable(params$survTable, row.names = T)

Feature selection and effect

params$effectPlot

params$binMap

params$contMap

Validation

If you performed a validation either at a cohort level (Validation tab) or at the individual level (patient tab), the results will be displayed in this section.

params$ValHist
params$ValKM

params$IndKM

AxelitoMartin/OncoCast documentation built on Dec. 13, 2020, 6:46 a.m.