function - 当函数与循环外的每个 out 一起正常工作时，在 while 循环中列出索引超出范围

Question

我正在构建一个while循环并遇到我的函数在while循环之外正常工作但在我的while循环中导致错误的问题。

这是while循环（仍在编码）：

diff=1

while (diff!=0):
  cents_new
  cents_old=cents_new.copy()
  print(cents_old)
  print(cents_new)
  feat_list=df_features.values.tolist()
  print(feat_list)
  dist_cent1=[]
  dist_cent2=[]
  dist_cent3=[]
  print(dist_cent1)
  print(dist_cent2)
  print(dist_cent3)
  dist_to_cent(cents_old, feat_list)
  print(dist_cent1)
  print(dist_cent2)
  print(dist_cent3)
  print(len(dist_cent1))
  print(len(dist_cent2))
  print(len(dist_cent3))
  min_index_list=[]
  print(min_index_list)
  min_assign(dist_cent1,dist_cent2,dist_cent3)
  print(min_index_list)
  print(len(min_index_list))
  df_features1=df_features.copy()
  df_features1['Closest Centroid']=min_index_list
  print(df_features)
  print(df_features1)
  feat_list1=df_features1.values.tolist()
  print(feat_list1)
  avg_list_cent1=[]
  avg_list_cent2=[]
  avg_list_cent3=[]
  print(avg_list_cent1)
  print(avg_list_cent2)
  print(avg_list_cent3)
  append_by_cent(feat_list1)
  print(avg_list_cent1)
  print(avg_list_cent2)
  print(avg_list_cent3)
  print(len(avg_list_cent1))
  print(len(avg_list_cent2))
  print(len(avg_list_cent3))
  cents_new=[]
  avg_and_assign(avg_list_cent1)
  avg_and_assign(avg_list_cent2)
  avg_and_assign(avg_list_cent3)



  
  



  diff=0

当我去运行 avg_and_assign 函数时，该函数具有以下代码：

def avg_and_assign(*lists):
    cents_new = []
    for lst in lists:
        alc = np.array(lst)
        alc_mean = np.mean(alc, axis=0)[:4]
        cents_new.append( np.ndarray.tolist(alc_mean) )
    return cents_new

cents_new = avg_and_assign(avg_list_cent1, avg_list_cent2, avg_list_cent3)

我收到以下错误：

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-104-2eb0f065bb1e> in <module>()
     16   min_index_list=[]
     17   #print(min_index_list)
---> 18   min_assign(dist_cent1,dist_cent2,dist_cent3)
     19   #print(min_index_list)
     20   df_features1=df_features[:]

<ipython-input-101-a817f672ae5a> in min_assign(l1, l2, l3)
     51    #Here, I make a list that will hold the same index from each of the three
     52    #centroid-to-feature distance lists.
---> 53    min_list=[l1[count], l2[count], l3[count]]
     54    #I then determine which index has the minimum value.
     55    min_index=min_list.index(min(min_list))

IndexError: list index out of range

这是引用另一个可以正常工作的函数，但现在在 while 循环中不是这样：

def min_assign(l1, l2, l3):
  #I chose to use a count to both orderly input indices and to drive a while 
  #loop to that in turn drives appendices to my list.
  count=0
  while count<150:
   #Here, I make a list that will hold the same index from each of the three
   #centroid-to-feature distance lists.
   min_list=[l1[count], l2[count], l3[count]]
   #I then determine which index has the minimum value.
   min_index=min_list.index(min(min_list))
   #I then append the index value+1 (because I don't have a 0-labeled centoroid)
   #to my list above the function
   min_index_list.append(min_index+1)
   #Increase the count because nobody likes an infinite loop. Then again, you're
   #a mathematician so you may know somebody who does, at least outside of Python.
   count+=1

#Here I run my function, being sure to input the centroid lists in numerical 
#order so that each index output by the function numerically coincides to the 
#centroid.
min_assign(dist_cent1,dist_cent2,dist_cent3)

我认为问题在于 while(count<150) 不再适用于 min_assign 不再有效。在 while 循环之前所有这些函数的第一次迭代中，每个列表都有 150 个条目。但是，150 个观测值中的每一个都应该与三个质心中的每一个都有距离，所以我不确定这会如何变化。但是，老实说，我曾多次尝试重做这些功能，但一遍又一遍地得到相同的错误。这是我运行的所有代码以供参考：

iris = pd.read_csv('Iris.csv')
iris=iris.iloc[:,1:]
iris.columns='sepal length', 'sepal width', 'petal length', 'petal width', 'species'
df=iris
df

df_features = df[['sepal length', 'sepal width', 'petal length', 'petal width']]
df_features

cent1=[5.4, 3.9,1.3,0.4]
cent2=[5.8,2.6,4.0,1.2]
cent3=[7.7,2.8,6.7,2.0]
cents=[cent1,cent2,cent3]
cents

df_cents=pd.DataFrame(cents, columns=['sepal length', 'sepal width', 'petal length', 'petal width'])
df_cents

cent_frame=pd.DataFrame(cents)
feat_list=df_features.values.tolist()

#Here I made lists to append the measured distances for each feature point from
#each centroid
dist_cent1=[]
dist_cent2=[]
dist_cent3=[]

#This is my distance function
def dist_to_cent(l1,l2):
  #I used a count to track which list would have the distance value appended.
  #There are 150 features, and each have to be measured against three different
  #datapoints, therefore there would be 450 measurements performed.
  count=0
  for i in l1:
    for j in l2:
      dist=((i[0]-j[0])**2 + (i[1]-j[1])**2 + (i[2]-j[2])**2 + (i[3]-j[3])**2)**0.5
      if count <=149:
        dist_cent1.append(dist)
      elif count<=299:
        dist_cent2.append(dist)
      else:
        dist_cent3.append(dist)
      count+=1


dist_to_cent(cents, feat_list)
#min_list= [dist_cent1[0],dist_cent2[0],dist_cent3[0]]

#print(min_list)

#print(min(min_list))

#print(min_list.index(min(min_list)))

#Now I have to determine the centroid with the minimum distance to each of the 
#features. I did this with the below function. I started by making a blank list
#to hold my minimum values.
min_index_list=[]
def min_assign(l1, l2, l3):
  #I chose to use a count to both orderly input indices and to drive a while 
  #loop to that in turn drives appendices to my list.
  count=0
  while count<150:
   #Here, I make a list that will hold the same index from each of the three
   #centroid-to-feature distance lists.
   min_list=[l1[count], l2[count], l3[count]]
   #I then determine which index has the minimum value.
   min_index=min_list.index(min(min_list))
   #I then append the index value+1 (because I don't have a 0-labeled centoroid)
   #to my list above the function
   min_index_list.append(min_index+1)
   #Increase the count because nobody likes an infinite loop. Then again, you're
   #a mathematician so you may know somebody who does, at least outside of Python.
   count+=1

#Here I run my function, being sure to input the centroid lists in numerical 
#order so that each index output by the function numerically coincides to the 
#centroid.
min_assign(dist_cent1,dist_cent2,dist_cent3)
min_index_list

#Finally, I append the list to my features dataframe with the below column 
#label.
df_features1=df_features.copy()
df_features1['Closest Centroid']=min_index_list

df_features
print(dist_cent1)
print(min_index_list)

feat_list1=df_features1.values.tolist()
#print(feat_list)

avg_list_cent1=[]
avg_list_cent2=[]
avg_list_cent3=[]

def append_by_cent(a_list):
 for i in a_list:
  if i[4]==1.0:
    avg_list_cent1.append(i)
  elif i[4]==2.0:
    avg_list_cent2.append(i)
  else:
    avg_list_cent3.append(i)


append_by_cent(feat_list1)

#I will be basing my averages off using numpy as below:
#a = numpy.array([[240, 240, 239],
#                 [250, 249, 237], 
#                 [242, 239, 237],
#                 [240, 234, 233]])
#print numpy.mean(a, axis=0)
#def avg_and_assign(*lists):
#    cents_new = []
#    for lst in lists:
#        alc = np.array(lst)
#        alc_mean = np.mean(alc, axis=0)[:4]
#        cents_new.append( np.ndarray.tolist(alc_mean) )
#    return cents_new
cents_new=[]
def avg_and_assign(l1):
  l1a=[i.pop(4) for i in l1]
  #print(l1)
  arrl1=array(l1)
  #print(arrl1)
  arrl1=average(arrl1, axis=0)
  new_cent1=arrl1.tolist()
  cents_new.append(new_cent1)


avg_and_assign(avg_list_cent1)
avg_and_assign(avg_list_cent2)
avg_and_assign(avg_list_cent3)




print(cents_new)

为什么会突然将 avg_and_assign 函数触发在 min_assign 函数中的错误，该函数在将前一个函数粘贴到循环之前运行得很好。这些函数如何在循环外以相同的顺序运行但在循环中中断？

我只是不明白这里发生了什么。

任何帮助都会很棒。

顺便说一句：我知道我的代码很烂，我知道 Python。请可怜我。

标签： functionwhile-loopiris-dataset