Question

kmeans = KMeans(n_clusters=5).fit(df)
print(kmeans.cluster_centers_.shape)

fig = plt.figure(figsize=(20,10))
for c in range(1,16):
fig.add_subplot(4,4,c).plot(kmeans.cluster_centers_[c-1,:])
plt.show()

I have an error which is 'index 5 is out of bounds for axis 0 with size 5'.

I think there is an problem on this numbers (n_clusters and figsize)

kmeans = KMeans(n_clusters=5).fit(df)

fig = plt.figure(figsize=(20,10))

How can i fix this?or change numbers?

Answer 1

Here in the given code error is in

for c in range(1,16):
fig.add_subplot(4,4,c).plot(kmeans.cluster_centers_[c-1,:])

Here

kmeans = KMeans(n_clusters=5).fit(df) will fit the input data in 5 number of clusters. So there will be 5 number cluster centers...

Now kmeans.cluster_centers_.shape[0] = 5... That means 5 centers with given data dimensions are created.

And the given snippet of code tries to access kmeans.cluster_centers_[5] , kmeans.cluster_centers_[6] ....kmeans.cluster_centers_[15] ... So it is giving error because only 5 centers are available...

for c in range(1,16):
fig.add_subplot(4,4,c).plot(kmeans.cluster_centers_[c-1,:])

Solution :-

Either change :-

kmeans = KMeans(n_clusters=5).fit(df) to

kmeans = KMeans(n_clusters=16).fit(df)

or

for c in range(1,16):
fig.add_subplot(4,4,c).plot(kmeans.cluster_centers_[c-1,:])

to

for c in range(1,5):
fig.add_subplot(3,2,c).plot(kmeans.cluster_centers_[c-1,:])