data(mtcars)
head(mtcars)
bst <- xgb.train(data = xgb.DMatrix(as.matrix(mtcars[, -11]), label = mtcars[, 11]),
                 objective = 'count:poisson', nrounds = 5)
pred <- predict(bst, as.matrix(mtcars[, -11]))
sqrt(mean((pred - mtcars[, 11]) ^ 2))