Bootstrap with replacement

[NeelaGitH]

1.okay so when we talk about bootstrapping with replacement, what's the significance of the word "replacement'?

2. And in bootstrapping when we split the data into different subsets to train different decision trees in random forest, suppose one row of data has already been used to train one decision tree, can that row be used again to train another decision tree? basically, can the subsets of data have some mutual rows?

3. and are the rows in the subsets of data called bootstrap samples? and when the data is split, are the no. of rows of data in each subset the same??

4. so here it says the 'max samples' default value is X.shape[0], so it basically says the max samples can be the total no. of rows in the dataframe?