sklearn的成對距離結果意外不對稱

我正在計算矢量元素之間的歐氏成對距離。我使用sklearn包中的pairwise_distances函數。然而，一些元素的結果矩陣僅僅是近似對稱的：在一個例子中，應該相等的元素的值僅等於小數點後15位數。sklearn的成對距離結果意外不對稱

我意識到這一點，因爲我在假定輸入矩陣對稱的下游分析中出現錯誤。我知道我可以將價值整理起來，但是造成這種情況的原因是什麼？

這裏是我試圖計算用於成對距離向量（它是一個大熊貓數據幀的一列）：

lag_measure_data[['bios_level']].values 

array([[ 0.76881030949999995538490793478558771312236785888671875 ], 
    [ 0.              ], 
    [ 0.67783090619999997183953155399649403989315032958984375 ], 
    [ 0.3228176074999999922710003374959342181682586669921875 ], 
    [ 0.75822395549999999087020796650904230773448944091796875 ], 
    [ 0.469808621599999975959605080788605846464633941650390625], 
    [ 0.989529862699999984698706612107343971729278564453125 ], 
    [ 0.              ], 
    [ 0.5575436799999999859522858969285152852535247802734375 ], 
    [ 0.9756440299999999954394525047973729670047760009765625 ], 
    [ 0.66511863289999995085821637985645793378353118896484375 ], 
    [ 0.978062709200000046649847718072123825550079345703125 ], 
    [ 0.473957179800000016900440868994337506592273712158203125], 
    [ 0.82409385540000001935112550199846737086772918701171875 ], 
    [ 0.56548685279999999497846374651999212801456451416015625 ], 
    [ 0.399505730399999980928527065771049819886684417724609375], 
    [ 0.474232963900000026313819034839980304241180419921875 ], 
    [ 0.34276307189999999369689476225175894796848297119140625 ], 
    [ 0.9985316859999999739017084721126593649387359619140625 ], 
    [ 0.9063241512999999915933813099400140345096588134765625 ], 
    [ 0.              ]])

這裏是我使用來獲得距離矩陣的命令：

d_matrix_lag = pairwise_distances(lag_measure_data[['bios_level']].values)

這裏我就不打印輸出距離矩陣，因爲它是太亂了，但作爲第一行中的例子爲第4列的值是

0.445992701999999907602756366 031826473772525787353515625

，而第4行和第一列的值是

0.4459927019999998520916051347739994525909423828125

來源

2017-01-16 user277194

什麼是向量提供這樣的距離？ – Dmitry

因此，如果您比較已打印的數組的第1個和第4個元素，則可以重現這些結果。 pairwise_distances（0.3228176074999999922710003374959342181682586669921875,0.76881030949999995538490793478558771312236785888671875）缺貨[700]：陣列（[[0.4459927019999998520916051347739994525909423828125]]） pairwise_distances（0.76881030949999995538490793478558771312236785888671875,0.3228176074999999922710003374959342181682586669921875）缺貨[701]：陣列（[[0.445992701999999907602756366031826473772525787353515625]]） – user277194

在上述我的註釋對兩個花車進行兩次計算，同時切換兩個參數的位置，結果略有不同 – user277194

我可以重現你的錯誤我的對稱性測試：

import numpy as np 

a = np.array([[ 0.76881030949999995538490793478558771312236785888671875 ], 
    [ 0.              ], 
    [ 0.67783090619999997183953155399649403989315032958984375 ], 
    [ 0.3228176074999999922710003374959342181682586669921875 ], 
    [ 0.75822395549999999087020796650904230773448944091796875 ], 
    [ 0.469808621599999975959605080788605846464633941650390625], 
    [ 0.989529862699999984698706612107343971729278564453125 ], 
    [ 0.              ], 
    [ 0.5575436799999999859522858969285152852535247802734375 ], 
    [ 0.9756440299999999954394525047973729670047760009765625 ], 
    [ 0.66511863289999995085821637985645793378353118896484375 ], 
    [ 0.978062709200000046649847718072123825550079345703125 ], 
    [ 0.473957179800000016900440868994337506592273712158203125], 
    [ 0.82409385540000001935112550199846737086772918701171875 ], 
    [ 0.56548685279999999497846374651999212801456451416015625 ], 
    [ 0.399505730399999980928527065771049819886684417724609375], 
    [ 0.474232963900000026313819034839980304241180419921875 ], 
    [ 0.34276307189999999369689476225175894796848297119140625 ], 
    [ 0.9985316859999999739017084721126593649387359619140625 ], 
    [ 0.9063241512999999915933813099400140345096588134765625 ], 
    [ 0.              ]]) 

from sklearn.metrics.pairwise import pairwise_distances 
dist_sklearn = pairwise_distances(a) 
print((dist_sklearn.transpose() == dist_sklearn).all())

越來越假作爲輸出。嘗試改用scipy.spatial.distance。您將獲得成對距離計算的距離向量，但可以將其與squareform轉化爲距離矩陣（）

from scipy.spatial.distance import pdist, squareform 

dist = pdist(a) 
sq = squareform(dist) 
print((sq.transpose() == sq).all())

這給了我對稱矩陣。希望這有助於

來源

2017-01-16 13:43:05 NKlink0r

謝謝，這節省了我的一天，因爲我對此感到頭疼。出於某種原因，我甚至在距離矩陣四捨五入時遇到困難。我開始失去了我對python的信仰...... – user277194

歡迎你。永遠不要對蟒蛇失去信心;） – NKlink0r

sklearn的成對距離結果意外不對稱

回答

相關問題