This question requires that you have a functioning implementation of the meanshift algorithm. Apply four iterations of the algorithm to the dataset 'galaxies_3D-short.xyz' (or '.csv') with the parameter k=10 and save the results to a new .xyz file. Then display the points in Meshlab.

How do the points distribute in space?

Build a dendrogram for languages.csv dataset.

This dataset includes several Germanic languages and contains lexical distances between them (more on these distances can be read here).

Read in the Wikipedia about language families of Germanic languages and choose which of them are clearly separated by the clustering algorithm.

(you can build a dendrogram from a distance matrix, see the TD)

Build a dendrogram for iris.csv (you may want to increase the size of the plot, see the hints in the TD).

Claim: There is a clear separation between the Iris-setosa cluster and all other irises.

Use the provided python script to analyze bluered.csv (22 two-dimensional points; this is the data described in the TD) and determine the minimal height in the dendrogram where 3 meaningful clusters can be seen.

Which is which?