workshop_contents.txt

Session 1: Dim. Red.& Clustering
1. DR::
   1. PCA
   2. ICA
   3. t-SNE
   4. MDS
1. Clustering
   1. K-means
   2. GMM
   3. DP
   4. Extension to overlapping clusters


Session 2: Regression & Classification
   1. Linear regr (assumptions: error model, variance, … )/Log. Regr
   2. Kernel method
   3. SVR/SVM
   4. Random Forest Reg/Class.
                
IK: I think we also need to consider parameter tuning / cross validation. Maybe that is something to do in the end of section 2? If so, the outline for the second section would be:
Session 2: Systematic validation + Regression & Classification
1. Regression
1. Linear regr (assumptions: error model, variance, … )/Log. Regr
   1. Kernel method
   2. SVR/SVM
   3. Random Forest Reg/Class.
1. Cross-validation
   1. Training / testing
   2. Cross-validation
   3. Parameter tuning
If we do this, we have a natural progression of exploratory analysis -> unsupervised -> supervised -> systematic validation.