Python 相关性';s转换为新的CSV文件
我有100个CSV文件:Python 相关性';s转换为新的CSV文件,python,pandas,csv,correlation,Python,Pandas,Csv,Correlation,我有100个CSV文件: Merge_Prediction_Groundtruth_Speed1.0_Buffer100.csv Merge_Prediction_Groundtruth_Speed1.0_Buffer200.csv Merge_Prediction_Groundtruth_Speed1.0_Buffer300.csv Merge_Prediction_Groundtruth_Speed2.0_Buffer100.csv Merge_Prediction_Groundtruth_
Merge_Prediction_Groundtruth_Speed1.0_Buffer100.csv
Merge_Prediction_Groundtruth_Speed1.0_Buffer200.csv
Merge_Prediction_Groundtruth_Speed1.0_Buffer300.csv
Merge_Prediction_Groundtruth_Speed2.0_Buffer100.csv
Merge_Prediction_Groundtruth_Speed2.0_Buffer200.csv
Merge_Prediction_Groundtruth_Speed2.0_Buffer300.csv
...............
所有CSV的结构数据如下所示:
BS Prediction Ground truth
BS1-BS1 0 0
BS1-BS2 0 2
BS1-BS3 2 35
BS1-BS4 0 0
BS1-BS5 0 0
BS1-BS6 0 2
BS1-BS7 0 0
BS1-BS8 0 2
BS1-BS9 0 0
BS2-BS1 0 1
...............
Files Correlation
Speed1.0_Buffer100 0.65
Speed1.0_Buffer200 0.51
Speed1.0_Buffer300 0.73
Speed2.0_Buffer100 0.36
Speed2.0_Buffer200 0.59
Speed2.0_Buffer300 0.44
...............
我想分析预测栏和地面真相栏之间的相关性。
我使用了以下代码:
df['Prediction'].corr(df['Ground truth'])
如果我一个接一个地分析,那就要花很长时间。
是否可以根据文件的最后一个标题一次性分析相关性并同时构建到一个CSV文件中。?
我的预期结果如下所示:
BS Prediction Ground truth
BS1-BS1 0 0
BS1-BS2 0 2
BS1-BS3 2 35
BS1-BS4 0 0
BS1-BS5 0 0
BS1-BS6 0 2
BS1-BS7 0 0
BS1-BS8 0 2
BS1-BS9 0 0
BS2-BS1 0 1
...............
Files Correlation
Speed1.0_Buffer100 0.65
Speed1.0_Buffer200 0.51
Speed1.0_Buffer300 0.73
Speed2.0_Buffer100 0.36
Speed2.0_Buffer200 0.59
Speed2.0_Buffer300 0.44
...............
提前感谢您。您可以在文件夹中读取csv文件
l=['Merge_Prediction_Groundtruth_Speed1.0_Buffer100.csv',
'Merge_Prediction_Groundtruth_Speed1.0_Buffer200.csv'
...]
比如:
然后使用concat
和groupby
pd.concat(d).groupby(level=0).apply(lambda x : x['Prediction'].corr(x['Groundtruth']))
您可以在文件夹中读取csv文件
l=['Merge_Prediction_Groundtruth_Speed1.0_Buffer100.csv',
'Merge_Prediction_Groundtruth_Speed1.0_Buffer200.csv'
...]
比如:
然后使用concat
和groupby
pd.concat(d).groupby(level=0).apply(lambda x : x['Prediction'].corr(x['Groundtruth']))
非常感谢。太棒了,谢谢你。这太神奇了。