Text clustering using k-Means and optimizing the amount of clusters
I have made the attached process for RapidMiner, which will do the following:
Take the text from different rows and cluster them by similarity, while also carrying over the unique id number of each row. An optimization of the number of clusters needed by the Elbow method is also part of the process.
It is probably not perfect, but it took me some time to scramble together from different sources.
I hope it might help people who are in the same situation as I was to get a better starting point.
Here is the file: