Can a human cluster the data-set manually?
http://i.stack.imgur.com/Ae6qa.png
Instead of using the clustering algorithms like Connectivity based clustering (hierarchical clustering),Centroid-based clustering,
Distribution-based clustering,
Density-based clustering etc
Can a Human manually cluster the Iris dataset?
For our convenience let us consider it as a two dimensional dataset.
By what means the human cluster the dataset? How will he/she cluster the dataset? Or the data clustering is done by the human?
http://i.stack.imgur.com/Ae6qa.png
Dear Parasha,
For a human observing variations in a simultaneous way for more than 2-3 varaibles is impossible. That is why multivariate analysis is soo important.
Regards.
Ludovic.
I guess it depends on the objective of the clustering. If you want to cluster a set of data points just to make groups you can do it manually. If you want to use the clustering result to work on new data not included in the original clustering, then you will need some kind of "definition" (model) of the clusters (e.g. distance from cluster center in the simplest case) to be able to assign new data points to the most likely cluster (classification).
There is a discussion on a similar topic for two-dimensional data sets in Chapter 2 of Everitt (2011), ``Cluster Analysis''. However, I agree with the previous comments. The probability of making an impartial judgement about homogeneity based on plots alone, especially when the dimension, p, exceeds 2 would be very low.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA