This function allows for calculating model prediction in a unified way.
# S3 method for surv_explainer
predict(object, newdata = NULL, output_type = "survival", times = NULL, ...)an explainer object - model preprocessed by the explain() function
data used for the prediction
character, either "risk", "survival" or "chf" depending on the desired output
a numeric vector of times for the survival and cumulative hazard function predictions to be evaluated at. If "output_type == "risk" this argument is ignored, if left NULL then it is extracted from object$times.
other arguments, currently ignored
A vector or matrix containing the prediction.
library(survival)
library(survex)
cph <- coxph(Surv(time, status) ~ ., data = veteran, model = TRUE, x = TRUE, y = TRUE)
rsf_ranger <- ranger::ranger(Surv(time, status) ~ .,
data = veteran,
respect.unordered.factors = TRUE,
num.trees = 100,
mtry = 3,
max.depth = 5
)
cph_exp <- explain(cph)
#> Preparation of a new explainer is initiated
#> -> model label : coxph ( default )
#> -> data : 137 rows 6 cols ( extracted from the model )
#> -> target variable : 137 values ( 128 events and 9 censored , censoring rate = 0.066 ) ( extracted from the model )
#> -> times : 50 unique time points , min = 1.5 , median survival time = 80 , max = 999
#> -> times : ( generated from y as uniformly distributed survival quantiles based on Kaplan-Meier estimator )
#> -> predict function : predict.coxph with type = 'risk' will be used ( default )
#> -> predict survival function : predictSurvProb.coxph will be used ( default )
#> -> predict cumulative hazard function : -log(predict_survival_function) will be used ( default )
#> -> model_info : package survival , ver. 3.7.0 , task survival ( default )
#> A new explainer has been created!
rsf_ranger_exp <- explain(rsf_ranger,
data = veteran[, -c(3, 4)],
y = Surv(veteran$time, veteran$status)
)
#> Preparation of a new explainer is initiated
#> -> model label : ranger ( default )
#> -> data : 137 rows 6 cols
#> -> target variable : 137 values ( 128 events and 9 censored )
#> -> times : 50 unique time points , min = 1.5 , median survival time = 80 , max = 999
#> -> times : ( generated from y as uniformly distributed survival quantiles based on Kaplan-Meier estimator )
#> -> predict function : sum over the predict_cumulative_hazard_function will be used ( default )
#> -> predict survival function : stepfun based on predict.ranger()$survival will be used ( default )
#> -> predict cumulative hazard function : stepfun based on predict.ranger()$chf will be used ( default )
#> -> model_info : package ranger , ver. 0.16.0 , task survival ( default )
#> A new explainer has been created!
predict(cph_exp, veteran[1, ], output_type = "survival")[, 1:10]
#> [1] 0.9959425 0.9896930 0.9831475 0.9739694 0.9692183 0.9619751 0.9570515
#> [8] 0.9494633 0.9414903 0.9358682
predict(cph_exp, veteran[1, ], output_type = "risk")
#> 1
#> 0.7354128
predict(rsf_ranger_exp, veteran[1, ], output_type = "chf")[, 1:10]
#> [1] 0.005945378 0.005945378 0.018315374 0.022121901 0.024215918 0.057908648
#> [7] 0.059175652 0.063733170 0.069066503 0.070342694