Package: familiar

Title: End-to-End Automated Machine Learning and Model Evaluation

Version: 1.4.3

Version: 1.4.4

Authors@R: c(

person("Alex", "Zwanenburg",

email = "alexander.zwanenburg@nct-dresden.de",

- Prevented an error that could occur when computing net benefit for decision curves of models that would predict class probabilities of exactly 1. This was a very rare error, as it would only occur if predicted class probabilities would have at most two distinct values, one of which is 1.0.

- Prevented an issue that could occur when computing linear calibration fits where the fit can be computed without residual errors. This would prevent the t-statistic and p-value from being correctly computed for binomial, multinomial and survival outcomes.

- Prevented an issue when computing linear calibration fits when all the expected values are the same. The model will then lack a slope. We now add a slope of 0, with an infinite confidence interval, if this is the case.

fit_coef<-stats::coef(fit)

no_slope<- is.na(fit_coef["expected"])&&data.table::uniqueN(calibration_data$expected)

if (no_slope)fit_coef["expected"]<-0.0

fit_conf_int<- suppressWarnings(stats::confint(fit))

if (no_slope)fit_conf_int["expected", ]<- c(-Inf,Inf)

fit_summary<- suppressWarnings(stats::summary.lm(fit))$coefficients

if (no_slope) {

fit_summary<- rbind(

fit_summary,

matrix(data= c(0.0,Inf,0.0,1.0),ncol=4)

)

rownames(fit_summary)<- c("(Intercept)","expected")

}

fit_summary[,3]<-fit_summary[,1]/fit_summary[,2]

fit_summary[,3][fit_summary[,2]==0.0]<-Inf

fit_summary[,4]<-2.0*stats::pt(abs(fit_summary[,3]),fit$df.residual,lower.tail=FALSE)

if(method=="linear"){

observed_interpolated<-stats::approx(

observed_interpolated<-suppressWarnings(stats::approx(

x=data$expected,

y=data$observed,

xout=expected_interpolated,

method="linear",

rule=1)$y

rule=1)$y)

}elseif(method=="loess"){

data[,":="("net_benefit"=n_true_positive/n-n_false_positive/n* (probability/ (1.0-probability)))]

data<-data[is.finite(net_benefit)]

if (nrow(data)<=1)return(NULL)

data<-data[,list("net_benefit"=min(net_benefit)),by="probability"]