K Means using elbow method

Question

How to use elbow method to find the best no of clusters to use K Means?

Answer 1

Elbow method allows the user to know the best fit number of clusters.

Follow the below steps:

1. Compute clustering algorithm (e.g., k-means clustering) for different values of k. For instance, by varying k from 1 to 10 clusters.
2. For each k, calculate the tot.withinss.
3. Plot the curve of above values against the number of clusters from step 1.
4. The value at the bend of the plot is considered as the best fit value for no of clusters.

totwss=sapply(1:10, function(k) { kmeans(mtcars$mpg,k)$tot.withinss})
k = data.frame(k = 1:6,totwss = totwss)
ggplot(k,aes(k,totwss))+geom_line()

In this case, you can take 2 or 3 as per your choice.