popcorn / xgb2cpp Goto Github PK

Hello,
thanks for the code, helped a lot!
I tried XGBRegressor but it constantly less around 0.5 than python predictions

I checked dump_raw.txt vs xgb_regressor.cpp numbers have same decimals, so it shouldn't be a problem

Do you have any idea why?

Hi Sir,

Does the code also workable for XGBRegression? Can you add a regression version?

Best Wishes,

Chen

your regular expression seems not to get the right feature, instead it uses the note id.

feature_index = re.search('[^f]', level[0]).group(0)

which should be:

feature_index = re.search('f(\d+)', level[0]).group(1)

to use the feature index from sample.

Dear Sir,
I am working on a project using xgboost, which I have to use c++ code. Your idea is really awesome. Thank you! I trained a model(classification of 0 and 1) with a medical dataset and tested with a piece of specific sample. I'v got a negative result(0). I dumped the model and used your code to generate a cpp code. Then I got two possibilities of 0.44 and 0.55. The results seem not the same if I understand it correctly.

From what I know, the model is a number of trees. And the sample goes through them. Normally I will get only one number? How does it give me two numbers, and possibilities based on them?

best wishes,
Jako

If I didn't miss something, there is no license specified for the code in this repository. Legally, this means that no one (except you, of course) is allowed to use it in any way.
As you put it on GitHub, I assume that this is not your intention. So to give users more certainty how they can use this script, it would be great if you could add some license. If you are not sure, which license to choose, this site may help: https://choosealicense.com