valdecy / c4.5 Goto Github PK

Hi,

Not a big issue, but just wanted to share that you might need to reset the dataframe indices (of Xdata and ydata) before entering them into dt_c45(). The indices caused the following error for me:

TypeError: '<' not supported between instances of 'float' and 'str'

referring to:

244 if len(np.unique(branch[i][0])) == 1 or len(branch[i]) == 1:

After resetting the indices, using df.reset_index(drop=True), the function works. Thanks!

this column "0" is actually the y which was added by the train function at first to the X dataframe, but mistakenly goes into the decision parameters.
with this mistake the training function runs just fine, but when predicting based on the tree we get an error that there is no column "0".

First of all, I apologize for not looking closely at your code, but both pre-pruning and post-pruning require validation sets for test pruning, but your Part 4 code does not have validation sets

I would like to ask that I changed the data in the data file by using the Iris data set in the UCI database.After the change, I ran the program and got the following error. You should be familiar with your own code, could you help me solve it?thank you

I mean the output of the script here is a list of IF-THEN rules cant we convert it to an image that contain the decision tree ?