search

amirsaffari / online-multiclass-lpboost

Online Multi-Class LPBoost and Gradient Boosting
http://www.ymer.org/amir/software/online-multiclass-lpboost/
MIT License
66 stars 33 forks source link

Samples assigned by fathernode are not updated in childnode's test #5

Open yiyusheng opened 6 years ago

yiyusheng commented 6 years ago

Hi,

In random forest, I find samples assigned by fathernode (depth=1) are not updated in childnode's (depth=2) test. Thus, those inherited samples stay at the childnode (depth=2) and are not passed to the child-childnode (depth=3) because they are not in the sample set of any test. I find this when I check the number of samples among all leaf node. The sum of sample numbers of all leaf nodes are not equals to the number of samples imported. I'd like to ask that whether I misunderstand your algorithm or it is a bug? Related codes are located at src/online_rf.cpp:75-85.

Thanks, Yi

amirsaffari commented 6 years ago

Hi Yi! Thanks. Would you be able to make a pull request that I can check.

yiyusheng commented 6 years ago

Hi Saffari,

Thank you for your reply. I haven't made a pull request yet since I am not sure whether it is my misunderstand or not. In my body of knowledge, random forest should assign every sample to leaf nodes. In your code, I find that samples from parent node stay in its childnode even it is not a leaf node and do not be passed to the child-childnode. Thus, I attempt to discuss with you for the correct understanding. Could you help me with that?