About clustering and regression…
jeudi 26 mars 2015, 13h30 - 14h30
We observe n points, and we want to make clusters to understand the
First, we will explain usual clustering methods.
Then, we will talk about regression, and we will define a mixture model in
The goal of this talk is to explain a methodology used to optimize
clustering in high-dimension, and explain quickly the theoretical results.
To conclude, we apply this methodology to real functional datasets: the
main example is electricity consumption curves. Our method is a way to
perform the understanding of such unsetteld dataset.