值错误:未知的标签类型:数组([0.11],..)在制作额外树模型时

Question

我试图在这个数据集上使用额外的树分类器，出于某种原因在

model.fit(trainx,trainy)

部分，它扔给我一个

ValueError: Unknown label type: array([[ 0.11],
       [ 0.12],
       [ 0.64],
       [ 0.83],
       [ 0.33],
       [ 0.72],
       [ 0.49],

错误。数组（[0.11] 是我的训练数据。我已经搜索了堆栈溢出，显然是由于 sklearn 无法识别数据类型，但我尝试了所有内容

trainy = np.asarray(trainy,dtype=float)
trainy=trainy.astype(float)

它不起作用，即使type（trainy）显示它的numpy.ndarray。谁能在这里指出我正确的方向？

代码如下：

import pandas as pd
import numpy as np
from sklearn.preprocessing import LabelEncoder
from sklearn import metrics
from sklearn.ensemble import ExtraTreesClassifier
from sklearn import cross_validation

def preProcess():
    df= pd.read_csv('C:/Users/X/Desktop/Managerial_and_Decision_Economics_2013_Video_Games_Dataset.csv',encoding ='ISO-8859-1')
    #drop non EA
    df = df[df['EA'] ==1]
    #change categorical variables
    le = LabelEncoder()
    nonnumeric_columns=['Console','Title','Publisher','Genre']
    for feature in nonnumeric_columns:
        df[feature] = le.fit_transform(df[feature])
    #set dataset and target variables
    dataset =df.ix[:, df.columns != 'US Sales (millions)']
    target = df['US Sales (millions)']
    trainx, testx, trainy, testy = cross_validation.train_test_split(
        dataset, target, test_size=0.3, random_state=0)
    #attempt to fix error?
    trainx=np.array(trainx)
    trainy = np.asarray(trainy, dtype="float")
    return trainx,testx,trainy,testy
def classifier():
    model =  ExtraTreesClassifier(n_estimators=250,
                              random_state=0)
    model.fit(trainx,trainy)
    return model.score(testx,testy)

trainx,testx,trainy,testy=preProcess()

我在python 3.5上使用scikit-learn 0.17

Answer 1

您的标签[[0.11], [ 0.12],.... .您应该使用ExtraTreesRegressor而不是ExtraTreesClassifier

从ForestClassifier的源代码：

 y : array-like, shape = [n_samples] or [n_samples, n_outputs]
            The target values (class labels in classification, real numbers in
            regression).

Answer 2

我的数组中有浮点数，创建one_hot时，我遇到了同样的错误。

training_labels = np.append(training_labels, [label])
...
y_one_hot = label_binarizer.fit_transform(training_labels)
ValueError: Unknown label type: (array([ 0. ,  0.1,

由于我正在做分类，我不得不将它们转换为字符串

training_labels = np.append(training_labels, [str(label)])
['0.0' '0.1' '-0.2' ..., '0.0' '0.0' '0.1']