Using XGBOOST in c++

Here is what you need：https://github.com/EmbolismSoil/xgboostpp

#include "xgboostpp.h"
#include <algorithm>
#include <iostream>

int main(int argc, const char* argv[])
{
    auto nsamples = 2;
    auto xgb = XGBoostPP(argv[1], 3); //特征列有4列, label有3个, iris例子中分别为三种类型的花，回归任何的话，这里nlabel=1即可

    //result = array([[9.9658281e-01, 2.4966884e-03, 9.2058454e-04],
    //       [9.9608469e-01, 2.4954407e-03, 1.4198524e-03]], dtype=float32)
    XGBoostPP::Matrix features(2, 4);
    features <<
        5.1, 3.5, 1.4, 0.2,
        4.9, 3.0, 1.4, 0.2;

    XGBoostPP::Matrix y;
    auto ret = xgb.predict(features, y);
    if (ret != 0){
        std::cout << "predict error" << std::endl;
    }

    std::cout << "intput : \n" << features << std::endl << "output: \n" << y << std::endl;
}

In case training in Python is okay and you only need to run the prediction in C++, there is a nice tool for generating static if/else-code from a trained model:

https://github.com/popcorn/xgb2cpp

I ended up using this after spending a day trying to load and use a xgboost model in C++ without success. The code generated by xgb2cpp was working instantly and also has the nice benefit that it does not have any dependencies.