Python 相关矩阵 - 只需要绝对值大于 .5 的列

Question

我有 41 个变量，其中大部分根本不相关。我只想包括几个列来说明相关性更高或负相关性更高的列。尽我所能，即使我看过很多文章和问题，我似乎也无法让它发挥作用。谢谢。

df.columns

索引(['ResponseId', 'Consent', 'AgeQualifier', 'Team', 'TeamOther', 'FanStrength', 'WinImportance', 'Emotion', 'Happiness', 'Satisfaction', 'Passion', 'ViewershipHomeGame', 'ViewershipRoadGame', 'ViewershipTVCable', 'ViewershipStreaming', 'ViewershipRestaurantBar', 'NameChangeViewershipHomeGame', 'NameChangeViewershipRoadGame', 'NameChangeViewershipTVCable', 'NameChangeViewershipStreaming', 'NameChangeViewershipRestaurantBar', 'Purchased', 'Purchased_Jersey_1', 'Purchased_Clothing_2', 'Purchased_Memorabilia_3', 'Purchased_Office_4', 'Purchased_Equipment_5', 'PurchaseIntentionNameChangeJersey', 'PurchaseIntentionNameChangeClothing', 'PurchaseIntentionNameChangeMemorabilia', 'PurchaseIntentionNameChangeHomeOffice', 'PurchaseIntentionNameChangeEquipment', 'Support_SeasonTickets', 'Support_Donations ', 'Support_Volunteer ', 'SupportNameChangeSeasonTickets', 'SupportNameChangeDonateMoney', 'SupportNameChangeVolunteer', 'State', 'Gender', 'Age', 'Ethnicity', 'EthnicityOther', 'Income', 'Drawing', 'Email'], dtype='object')

correlation_matrix = df.corr().round(2)

无花果, 斧头 = plt.subplots(figsize=(50,50)) sns.heatmap(data=correlation_matrix,cmap = 'rainbow', annot=True, ax=ax)

想法？

Answer 1

清理后的矩阵为

	Consent	AgeQualifier	Team	FanStrength	WinImportance
Consent	NaN	NaN	NaN	NaN	NaN
AgeQualifier	NaN	1.0	NaN	NaN	NaN
Team	NaN	NaN	1.00	0.02	0.02
FanStrength	NaN	NaN	0.02	1.00	0.69
WinImportance	NaN	NaN	0.02	0.69	1.00

要解决此问题，您需要 select 任何矩阵值非对角线且绝对值 >0.5

temp = df[(df>0.5)&(df!=1)].abs().max()
print(temp[~temp.isna()])

这将产生在相关矩阵中至少有一个相关性 >0.5

的列名

这会产生

FanStrength      0.69
WinImportance    0.69
dtype: float64

Python 相关矩阵 - 只需要绝对值大于 .5 的列

Python Correlation Matrix - Only Want columns that have absolute value more than .5

python

correlation