search for: truepos

...<- cforest(as.factor(remitter)~., data=cv.train, control=cforest_control(mtry=mtry, ntree=ntrees, mincriterion=0.95)) cv.evaluate$prediction <- predict(cv.model, newdata=cv.evaluate) obj.pred<-predict(cv.model) pred.table<-table(cv.evaluate$prediction, cv.evaluate$remitter) truepos<-pred.table[1,1] trueneg<-pred.table[2,2] falsepos<-pred.table[2,1] falseneg<-pred.table[1,2] sumvar<-epi.tests(c(truepos, falsepos , falseneg, trueneg), verbose=T) epi.tests(c(truepos, falsepos , falseneg, trueneg), verbose=F) # Calculate the overall accuracy. cv.eva...

...y used. Again, the internal integrate function yields: > integrate(ff, lower=target, upper=Inf) 0.7580801 with absolute error < 7.2e-05 Now, my understanding of the change of variable needed for Gauss-Laguerre in this instance is simple, x = y_i + target. As such, the integration should be truePos <- function(target, m, mu, sigma, sigma_i, Q = 30){ gq <- gauss.quad(Q, kind="laguerre") nodes <- gq$nodes whts <- gq$weights y <- pnorm((nodes + target - m)/sigma_i) * dnorm(nodes + target, mu, sigma) sum(y * exp(nodes)* whts) } > truePos(target = 200,...