Question

我正在使用 0 ，如下所述，对二进制分类使用交叉验证（类标签为1和from sklearn.ensemble import RandomForestClassifier from sklearn.model_selection import StratifiedKFold, cross_val_score clf=RandomForestClassifier(random_state = 42, class_weight="balanced") k_fold = StratifiedKFold(n_splits=10, shuffle=True, random_state=0) accuracy = cross_val_score(clf, X, y, cv=k_fold, scoring = 'accuracy') print("Accuracy: " + str(round(100*accuracy.mean(), 2)) + "%") f1 = cross_val_score(clf, X, y, cv=k_fold, scoring = 'f1_weighted') print("F Measure: " + str(round(100*f1.mean(), 2)) + "%")）。

现在，我想使用cross validation类并具有pred = clf.predict_proba(X)[:,1] print(pred) probs = clf.predict_proba(X) best_n = np.argsort(probs, axis=1)[:,-6:]个结果的预测概率对数据进行排序。为此，我尝试了以下两种方法。

    #define MES_LEN 100    
    #define m 23    
    typedef struct {    
       char index[m];    
       char name[MES_LEN];    
       char release[MES_LEN];    
       char length[MES_LEN];    
       char artist[MES_LEN];    
       } INFORM;    
    typedef struct list_elem {    
       INFORM inform;    
       struct list_elem *next;    
    }    
    LEL;    
    LEL* list;    
    void AddElement(void);    // Writing new element to file (DATAinput)
    void LoadData(void); // Reading previously written data    
    void ReadElement (void); // Reading data from existing file data (DATA)    
    void OutpuAll(void); // Output in console all previously added data from both files    
    int menu(void);    
    int main() {    
    ...                 
            case 1: OutpuAll(); break;                       // Output all data    
            case 2: AddElement(); ReadElement(); break;    // Add element to file                 
            case 3: LoadData(); break;             // Loading data from the file    
    ...
    }    
        void AddElement (void) // I have an empty file that i fill and then read and link to my main structure(this works fine)    
        {           
        char albN[25], albR[11], albL[11], albA[25], albI[5];    
        FILE * fileD;    
        fileD = fopen("DATAinput.txt", "r");    
    ...      
        printf("Enter album artist \n");    
            scanf("%s", albA);    
            fprintf(fileD, "%s ", albA);    
            fclose(fileD);    

    } 

    void ReadElement (void) // Reading previously written data(AddElement)    
    {    
        LEL *ptr;    
        ptr = (LEL*)malloc(sizeof(LEL));    
        FILE * fileD;    
        fileD = fopen("DATAinput.txt", "r");    
        if (fileD == NULL) {    
            printf("Can not open the file\n");    
            exit(1);    
        }    
        while(fscanf (fileD, "%s %s %s %s %s", ptr->inform.index, ptr->inform.name, ptr->inform.release, ptr->inform.length, ptr->inform.artist) != EOF)    
        { ptr->next=NULL;    
            printf("\n Id:%s Album: %s  Released: %s  Length: %s  Artist: %s", ptr->inform.index, ptr->inform.name, ptr->inform.release, ptr->inform.length, ptr->inform.artist); }    
        fclose(fileD);    
        free(ptr);    
    }    
    void LoadData(void)

 // Reading existing data from file    

    {    
        LEL *current, *head;    
        LEL *ptr;    
        ptr = (LEL*)malloc(sizeof(LEL));     
        FILE * file;     
        file = fopen("DATA.txt", "r");    
        if (file == NULL) {    
            printf("Can not open the file\n");    
        }    
        while(fscanf (file, "%s %s %s %s %s", ptr->inform.index, ptr->inform.name, ptr->inform.release, ptr->inform.length, ptr->inform.artist) != EOF)    
        {  printf("\n %s Album: %s  Released: %s  Length: %s  Artist: %s", ptr->inform.index, ptr->inform.name, ptr->inform.release, ptr->inform.length, ptr->inform.artist); }       
               fclose(file);    
               ptr->next = NULL;    
        }    
    void OutpuAll(void) // Trying to output data that i previously loaded from files to structure.    
    {    
        LEL *ptr; 
LEL* ptr = list;
    while (ptr!=NULL) {  
            printf("\n Id:%s Album: %s  Released: %s  Length: %s  Artist: %s", ptr->inform.index, ptr->inform.name, ptr->inform.release, ptr->inform.length, ptr->inform.artist);}     
       }
    ptr = ptr->next;
}

我收到以下错误

NotFittedError ：不适合此RandomForestClassifier实例然而。使用此方法之前，请使用适当的参数调用“适合”。

对于这两种情况。

我只是想知道我在哪里弄错了。

很高兴在需要时提供更多详细信息。

Answer 1

我使用以下代码解决了我的问题：

proba = cross_val_predict(clf, X, y, cv=k_fold, method='predict_proba')
print(proba[:,1])
print(np.argsort(proba[:,1]))

Answer 2

看看documentation，它指定了概率是根据树的平均结果来计算的。

在您的情况下，您首先需要调用fit()方法以在模型中生成发束。使模型适合训练数据后，您可以调用predict_proba()方法。

这也在错误中指定。

# Fit model
model = RandomForestClassifier(...)
model.fit(X_train, Y_train)

# Probabilty
model.predict_proba(X)[:,1]

Answer 3

如果您想将CV模型用于看不见的数据点，请使用以下方法。

from sklearn import datasets
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import cross_validate

iris = datasets.load_iris()
X = iris.data
y = iris.target
clf = RandomForestClassifier(n_estimators=10, random_state = 42, class_weight="balanced")

cv_results = cross_validate(clf, X, y, cv=3, return_estimator=True)

clf_fold_0 = cv_results['estimator'][0]

clf_fold_0.predict_proba([iris.data[133]])

# array([[0. , 0.5, 0.5]])

如何在scikit-learn中使用交叉验证来获得预测概率

3 个答案: