在熊猫中将过采样的数据集另存为csv文件

时间:2020-08-24 07:59:31

标签: python pandas numpy resampling smote

我是Python的新手,如果太简单了,请提前道歉。找不到任何内容,this question没有帮助。

我的代码是

# Split data
y = starbucks_smote.iloc[:, -1]
X = starbucks_smote.drop('label', axis = 1)

# Count labels by type
counter = Counter(y)
print(counter)
Counter({0: 9634, 1: 2895})

# Transform the dataset
oversample = SMOTE()
X, y = oversample.fit_resample(X, y)

# Print the oversampled dataset
counter = Counter(y)
print(counter)
Counter({0: 9634, 1: 9634})

如何保存过采样的数据集以备将来使用?

我尝试过

data_res = np.concatenate((X, y), axis = 1)
data_res.to_csv('sample_smote.csv')

遇到错误

ValueError: all the input arrays must have same number of dimensions, 
but the array at index 0 has 2 dimension(s) and the array at index 1 has 1 dimension(s)

感谢任何提示!

1 个答案:

答案 0 :(得分:2)

您可以创建数据框:

data_res = pd.DataFrame(X)
data_res['y'] = y

,然后将data_res保存到CSV。

基于级联numpy.arrays的解决方案也是可行的,但是需要np.vstack才能使尺寸兼容:

data_res = np.concatenate((X, np.vstack(y)), axis = 1)
data_res = pd.DataFrame(data_res)