错误总结：OneHotEncoder

OneHotEncoder 中
categorical_features代表类别属性的索引数值，n_values代表categorical_features中每个属性含有多少个类别
toarray()是集合转数组的意思

OnehotEncoder在实际应用中的理解： https://blog.csdn.net/qq_24509229/article/details/80183173

正确代码：
dataset = pd.read_csv(‘50_Startups.csv’)
X = dataset.iloc[ : , :-2].values
Y = dataset.iloc[ : , 3 ].values # 得到一维的行向量 Y_shape (50,) Y [‘New York’ ‘California’ ‘Florida’ ‘New York’ ‘Florida’ ‘New York’…]

from sklearn.preprocessing import LabelEncoder, OneHotEncoder
labelencoder = LabelEncoder()
Y = labelencoder.fit_transform(Y) # 把[‘New York’ ‘California’ ‘Florida’ ‘New York’ ‘Florida’ ‘New York’…] 转为 [2 0 1 2 1 2…]

onehotencoder = OneHotEncoder(categorical_features = [0]) # 取Y 的第0列
Y = Y.reshape(-1,1) # 等同于Y.shape (50, 1) 将 1行50列的[2 0 1 2 1 2…] 转化为 50行1列的 [[2] [0] [1] [2]…]

onehotencoder.fit(Y)
Y = onehotencoder.transform(Y).toarray()
print(‘Y’,Y)
print(‘Y.shape’,Y.shape)

最终得到：
[[0. 0. 1.]
[1. 0. 0.]
[0. 1. 0.]
[0. 0. 1.]