ValueError異常：未知的不支持多類因變量

Question

我試圖適應sklearn載體，但我收到此錯誤：

ValueError: unknown is not supported This is my code:

X = df_features.values 
    X = X.reshape((len(X),len(df_features.columns))) 
    Y = df_train['action'].values 
    Y = Y.reshape((len(Y),)) 

pipeline = Pipeline([ 
('clf', RandomForestClassifier()) 
]) 

parameters = { 
    'clf__max_depth': [5,7,9], 
    'clf__max_features': [3,4,5], 
    'clf__min_samples_leaf': [3,4,5,6,7], 
    'clf__bootstrap': [True] 
} 

score_func = make_scorer(metrics.f1_score,average='weighted') 

grid_search = GridSearchCV(pipeline, parameters, n_jobs=3, 
    verbose=1, scoring=score_func) 

grid_search.fit(X, Y) 

這是Y採樣數據：

['NOTHING', 'NOTHING', 'SELL', 'SELL', 'NOTHING', 'NOTHING', 'NOTHING']

我該如何解決這個問題？
謝謝

Answer 1

請檢查x和y的類型和大小。另外，您是否有足夠的樣本來獲取所需的max_depth和min_samples_leaf？

以下示例似乎正常工作。我使用了虹膜數據，並將一個交叉驗證作爲一個例子。

from sklearn.pipeline import Pipeline 
from sklearn.model_selection import GridSearchCV 
from sklearn.metrics import fbeta_score, make_scorer 
from sklearn.ensemble import RandomForestClassifier 
from sklearn.datasets import load_iris 
import numpy as np 
from sklearn import metrics 
from sklearn.model_selection import LeaveOneOut 


loo= LeaveOneOut() 
data = load_iris() 

x = data.data 
x = x[0:14,:] 
x.shape 

y = ['NOTHING', 'NOTHING', 'SELL', 'SELL', 'NOTHING', 'NOTHING','SELL','SELL','NOTHING','SELL','SELL','NOTHING','NOTHING','NOTHING'] 
y = np.asarray(y) 
y = y.reshape(14,1) 
y = y.astype('str') 


pipeline = Pipeline([ ('clf', RandomForestClassifier())]) 

parameters = {'clf__max_depth': [1,2,3], 'clf__max_features': [1,2,3], 'clf__min_samples_leaf': [1,2,3], 'clf__bootstrap': [True] } 

score_func = make_scorer(metrics.f1_score,average='weighted') 

grid_search = GridSearchCV(pipeline, parameters, n_jobs=1 , verbose=1, scoring=score_func, cv = loo) 

grid_search.fit(x, y) 

結果

Fitting 14 folds for each of 45 candidates, totalling 630 fits 
[Parallel(n_jobs=1)]: Done 630 out of 630 | elapsed: 33.7s finished 

希望這有助於