I am relatively a learner in the field of datascience. Recently I came across these concepts and I am really keen to implement them - i.e. the concept of multimodal clustering applications. (From here I got the idea - https://scikit-learn.org/stable/modules/biclustering.html)
I know about different clustering algorithms like DBSCAN, OPTICS, K-Means (which is very popular), etc. I understand that in all these algorithms, a single column data-points are considered for clustering a dataset.
Suppose someone has a dataset like: http://archive.ics.uci.edu/ml/datasets/Iris
How can one use 3 or more columns to cluster the distinct classes in this Iris dataset i.e. as per the defined terms how to implement multi-modal classification on this kind of dataset.
Or being new - I doubt that I am confusing it with multi-dimensional cube concepts. It would be a great help if someone could clarify and explain this to me.