sklearn-porter:python训练模型供其他语言调用

sklearn-porter源码地址：https://github.com/nok/sklearn-porter

安装指令：

pip install sklearn-porter

部分windows系统会报错，原因在于编码，转化一下编码方式。
先列出目前所支持的机器学习的模型到转换语言的对应关系吧:

官方提供的例程：

from sklearn.datasets import load_iris
from sklearn.tree import tree
from sklearn_porter import Porter

# Load data and train the classifier:
samples = load_iris()
X, y = samples.data, samples.target
clf = tree.DecisionTreeClassifier()
clf.fit(X, y)

# Export:
porter = Porter(clf, language='java')#这里可以改变转化语言
output = porter.export(embed_data=True)
print(output)

转化后的语言为java，这里我转化为C语言：

#include <stdlib.h>
#include <stdio.h>
#include <math.h>

int predict(float features[4]) {

    int classes[3];
        
    if (features[3] <= 0.800000011920929) {
        classes[0] = 50; 
        classes[1] = 0; 
        classes[2] = 0; 
    } else {
        if (features[3] <= 1.75) {
            if (features[2] <= 4.950000047683716) {
                if (features[3] <= 1.6500000357627869) {
                    classes[0] = 0; 
                    classes[1] = 47; 
                    classes[2] = 0; 
                } else {
                    classes[0] = 0; 
                    classes[1] = 0; 
                    classes[2] = 1; 
                }
            } else {
                if (features[3] <= 1.550000011920929) {
                    classes[0] = 0; 
                    classes[1] = 0; 
                    classes[2] = 3; 
                } else {
                    if (features[2] <= 5.450000047683716) {
                        classes[0] = 0; 
                        classes[1] = 2; 
                        classes[2] = 0; 
                    } else {
                        classes[0] = 0; 
                        classes[1] = 0; 
                        classes[2] = 1; 
                    }
                }
            }
        } else {
            if (features[2] <= 4.8500001430511475) {
                if (features[1] <= 3.100000023841858) {
                    classes[0] = 0; 
                    classes[1] = 0; 
                    classes[2] = 2; 
                } else {
                    classes[0] = 0; 
                    classes[1] = 1; 
                    classes[2] = 0; 
                }
            } else {
                classes[0] = 0; 
                classes[1] = 0; 
                classes[2] = 43; 
            }
        }
    }

    int index = 0;
    for (int i = 0; i < 3; i++) {
        index = classes[i] > classes[index] ? i : index;
    }
    return index;
}

int main(int argc, const char * argv[]) {

    /* Features: */
    double features[argc-1];
    int i;
    for (i = 1; i < argc; i++) {
        features[i-1] = atof(argv[i]);
    }

    /* Prediction: */
    printf("%d", predict(features));
    return 0;

}

可以看到，sklearn-porter的功能就是把模型的预测逻辑以代码的形式表现出来，如果想向在python中调用joblib存取模型还需要再进行进一步的修改。