Due to recent updates, all users are required to create an Altair One account to login to the RapidMiner community. Click the Register button to create your account using the same email that you have previously used to login to the RapidMiner community. This will ensure that any previously created content will be synced to your Altair One account. Once you login, you will be asked to provide a username that identifies you to other Community users. Email us at Community with questions.
When k = 2, 3, 4?
Contributor II
Below is my customer segmentation data which I ran in AutoModel.
Below are screen shots when k = 2, 3 and 4. How can I tell which k is best?
I do not have access to the elbow method or silhouette analysis.
I looked at the three Davis-Bouldin indices which measure 5.415, 3.666 and 4.121.
Wikepedia calls this an internal evaluation scheme, where the validation of how well the clustering has been done is made using quantities and features inherent to the dataset. Due to the way it is defined, as a function of the ratio of the within cluster scatter, to the between cluster separation, a lower value will mean that the clustering is better.
Should I assume the index 3.666 means k = 3 is better?
Thanks for your time.
Tony
Below are screen shots when k = 2, 3 and 4. How can I tell which k is best?
I do not have access to the elbow method or silhouette analysis.
I looked at the three Davis-Bouldin indices which measure 5.415, 3.666 and 4.121.
Wikepedia calls this an internal evaluation scheme, where the validation of how well the clustering has been done is made using quantities and features inherent to the dataset. Due to the way it is defined, as a function of the ratio of the within cluster scatter, to the between cluster separation, a lower value will mean that the clustering is better.
Should I assume the index 3.666 means k = 3 is better?
Thanks for your time.
Tony
0
Best Answer
-
jacobcybulski
Member, University Professor Posts: 391 
Unicorn
The AutoModel does not perform cluster optimisation for k-Means, so if you run several experiments, the best distributed cluster model is the one which gives Davis-Bouldin measure closest to zero. However, if you select x-Means clustering it will return the optimum cluster number in between the specified range between minimum and maximum.0
Answers