I am a beginner in Knime, so sorry for the newbie Question
I have a project in which I have to create a undifined number of clusters.
What is the best solution?
Should I use something like a k-mean and try a few solutions with different k or is there any possibility to use another clustering method? maybe a plugin?
you may want to find a number of clusters that fits best to your data.
Therefore you can for example calculate an error value for the kmeans results. You may also get an overview by taking a look to the dendrogram of the hirarchical clustering.
if you want to use additional method (for example density based things) you can install the weka plugin. there is also a kmeans implementation that finds out the number of cluster by itself (with an errorĀ value). It's called x means.
All clustering methods have their advantages and drawbacks and fit different for different data, distributions etc.
I would suggest you first take a deeper look in the clustering literature.