teaching/enseignement

Machine Learning


See also Course material, Other LABS, Course projects

LABS

3. Unsupervized Learning

Context
Consider as an example data consisting of the composition of 45 pottery found in Britain dating back to Roman time. The composition is the content of 9 components (iron, manganese ....). They come from 5 different ovens. The data matrix is ​​the 45 × 9 matrix of the compositions.

Goal
Use the Kmeans and Gaussian mixture algorihms to cluster the poteries dataset.

Example : examples of codes are provided for the digits data.


Questions
1. Run the Kmeans algorithm, compare the obtained classificaiton to the four labels using a confusion matrix
2. Run the GMM algorithm , compare the obtained classificaiton to the four labels using a confusion matrix
3. Compare the clustering results to each other (compute the number of well classified examples).

Codes: Poteries_clustering_ToStart.py, Poteries_clustering_ToStart.py, Digits_clustering.py, Digits_clustering.R