CODEWITHSUNDEEP
rcatboost
akshay
akshay

Reputation: 1

Catboost predict probabilties are negative

I'm new to catboost and i'm trying to run the algorithm on the iris dataset in R on a single iteration. I noticed negative probabilities from the output of predict function. Any idea why this happens?. Also, is there a way to plot a catboost tree in R?

test <- as.data.frame(iris)
test$y <- ifelse(test$Species=="setosa",1,0)
x_iris <- test[,c("Sepal.Length","Sepal.Width","Petal.Length","Petal.Width")]
y_iris <- test[,"y"]

train <- as.data.frame(lapply(x_iris, as.numeric))
train_pool <- catboost.load_pool(data = train, label = y_iris)

model <- catboost.train(train_pool,  NULL,
                        params = list(loss_function = 'Logloss',
                                      iterations = 1, metric_period=10))
print(prediction)

Upvotes: 0

Views: 692

Answers (1)

Suman Pal
Suman Pal

Reputation: 11

In python: we have to set prediction_type='Probability'

Upvotes: 1

Related Questions