plot.mtlr: Graphical Representation of Feature Weights

In MTLR: Survival Prediction with Multi-Task Logistic Regression

Description

Plot the weights of an mtlr object. If packages ggplot2 and reshape2 are not installed, a bargraph of feature influence is given where influence is defined as the sum of absolute values of the feature weights across time. If ggplot2 and reshape2 are installed then a plot of feature weight across time is given.

Usage

## S3 method for class 'mtlr'
plot(x, numfeatures = 5, featurenames = c(), digits,
  ...)

Arguments

x	an object of class mtlr (result from calling mtlr).
numfeatures	the number of weight to plot. Default is 5. The most influential features are chosen first.
featurenames	the names of the specific weight to plot. These should correspond to the names in x$weight_matrix. If featurenames are supplied, then numfeatures is ignored.
digits	the number of digits to round to for the value of the time points.
...	for future methods

Examples

#These examples are geared towards users who have installed ggplot2 and reshape2.
library(survival)
mod <- mtlr(Surv(time,status)~., data = lung)
#Basic plot with 5 most influential features
plot(mod)
#Plot all 8 features
plot(mod, numfeatures = 8)
#Suppose we want to see specifically the "meal.cal" and "ph.karno" features:
plot(mod, featurenames = c("meal.cal", "ph.karno"))

MTLR documentation built on June 4, 2019, 1:02 a.m.