Random Forest Predict

Question

I have a training dataset which has 40,000 rows and I was able to successfully generate a randomForest for my dataset.

I am trying to now use it to predict on my test set. My training and test data sets are 2 different data frames and the columns names between the 2 data frames do not match. I am getting an an error in when I run my predict.

library(randomForest)
set.seed(2018)
new_train_rf= randomForest(workdf.V1~.,data = new_train_df, mtry=6, ntree=25)
new_train_rf
summary(new_train_rf)

Test Dataset predictions:

test_pred = predict(new_train_rf, newdata=new_test_df)
test_pred
summary(test_pred)

Error in eval(predvars, data, env) : object 'Var57' not found

Column Names in Test Dataframe:

testdf.Var218_UYBR, testdf2.Var6, testdf2.Var13, testdf2.Var21

Column Names in Training Datafame:

workdf.Var218_UYBR, tempdf2.Var6, tempdf2.Var13, tempdf2.Var21

Please help! I am new to R and I have been trying to figure out why my prediction is not working

Pretty much any model expects the same data frame structure for training and prediction, you cannot have different ones. — user2974951

Sindhu Viswanathan Sindhu Viswanathan · Accepted Answer · 2018-12-10T19:48:22

I had my test and training data split into 2 different data frames. I was able to run my predictions successfully, after I renamed the columns in my df's (test and training) to match.

Random Forest Predict

1 Answers