Bruger:  


News

Ready for the 2017 version

This website is now updated for the 2017 course start.

Podcast links

Module Podcast (Links will be updated throughout the course)
1 Welcome and practical info (13 min)
1 Introduction to Chemometrics, part 1 (20 min)
1 Introduction to Chemometrics, part 2 (18 min)
1 Food example. PBB example of Sensory (and other) data on frozen peas. (12 min)
1 QSAR example: Permeability discrimination of drug candidates. (8 min)
1 Lattin Chapter2. A little recap of linear algebra(matrices). (20 min)
1 Correlation and covariance. Link between multivariate data and simple correlations. (5 min)
   
2 PCA, Part 1 by JCF (25 min)
2 PCA, Part 2 by JCF (35 min)
   
2 PCA intro by Per BB (course 02935)(21 min)
2 Per BB: PCA on Iris data, course 02593 (31 min)
2 Per BB: PCA on spectral data, course 02593 (10 min)
   
3 R-tutorial: Exploring the Leslie Salt data (20 min)
3 Multiple Linear Regression, MLR. Basics: The model and approach.(21 min)
3 Multiple Linear Regression, MLR. Computations, Statistical inference and interpretation. (22 min)
3 MLR. Model validation. Start. (1.5 min)
3 MLR. Model validation. Residual plotting and modelling non-linearities and interactions. (12 min)
3 Multicollinearity. And a perspective towards the dimension reduction techniques later in the course (3.5 min)
3 R-tutorial: MLR in R. Running MLR, stepwise variable elimination and statistical inference. (16 min)
3 R-tutorial. MLR in R. Model Validation: residual plotting and modelling non-linerarities and interactions. Predicting new observations in R. (21 min)
   
4 What is Principal Component Analysis. A brief initial ' what is PCR?'(3.5 min)
4 Motivating example 1. See how PCR makes better sense than MLR when the x'es are highly correlated.(14 min)
4 Spectral data motivation. A recap of PCA of spectral type data.(7.5 min)
4 What is PCR then. A final wrap up of this.(2 min)
4 PCR - how to do it? A brief overview of the analysis approach.(3 min)
4 Cross validation and friends. Cross validation and other validation and resampling techniques. Model selection. Bias and variance.(28 min)
4 R Tutorial, PCR. See how to use the 'pcr' function of the 'pls' package. (35 min)
   
5 Correspondence Analysis (26 min)
   
6 PLS R motivation(13 min)
6 PLS, what is it? (14 min)
6 PLS - how to do it! (13 min)
6 PLS - what is it - a chemometrics perspective. (12 min)
6 Rtutorial PLS, yarn data (21 min)
   
7 Optimal predictor? Finding optimal predictor by CV (cross validation): Bias, variance and model complexity. (15 min)
7 MLR and the problem. MLR produces solutions with too large coefficients. (10 min)
7 Ridge Regression, Definition (7.5 min)
7 Lasso Regression, Definition.(2 min)
7 Ridge Regression. The solution (3.5 min)
7 Ridge and Lasso. L2 and L1 regression methods - an overview. (12 min)
7 Ridge Regression . A small example. (3.5 min)
7 Ridge Regression. Question to small example. (2.5 min)
7 Ridge and Lasso. How to analyze data by Ridge and Lasso Regression - the overall approach. (3.5 min)
7 RTutorial, Ridge Regression. (15 min)
7 RTutorial, Lasso Regression. (7.5 min)
7 RTutorial, Ridge and Lasso Regression. How to predict in R.(3 min)
   
8 Discriminant Analysis: An introduction (14 min)
8 Discriminant Analysis . Different priors and miss classification costs (9 min)
8 Discriminant Analysis (17 min)
8 Discriminant Analysis: Fisher and CVA. (10 min)
8 Discriminant Analysis. Relations between Bayes, LDA, QDA and CVA. (4 min)
8 Discriminant Analysis . k-Nearest Neighbour (7.5 min)
8 Discriminant Analysis. Regression and PLS (5 min)
8 Discriminant Analysis. Regression, PLS and summary. (6.5 min)
8 RTutorial Discriminant Analysis, Iris data example. Various methods, part 1.(17 min)
8 RTutorial Discriminant Analysis, Iris data example. Various methods, part 2.(10 min)
   
9 Søren Havelund Welling(SHW): Supervised learning and classical statistics (4 min)
9 Random Forests: The Tree model, bagging and random forests. (19 min)
9 R-tutorial (SHW): Random forests. How to run Random forests in R. An example. (8 min)
9 R-tutorial, part 2 (SHW). Random Forests in R.(15 min)
9 Random Forests. Learning from the model: Interpretation. Using ForestFloor - an R-package developed by SHW. (17 min)
9 R-tutorial - exercise data (SHW). Presenting the two data sets for the exercises.(15 min)
   
10 Cluster Analysis. Lecture by Jens Chr. Frisvad on Cluster Analysis.(24 min)
   
12 Multidimensional scaling_PCO_MDS_NMS in general (26 min)
   
13 Overview Lecture 1. Overview of classification in Chemometrics, by Jens Chr. Frisvad. (14 min)
   
13 Overview Lecture 2A: Regression and discrimination, by Per B.B. (17 min)
13 Overview Lecture 2B. Bias and Variance in prediction, optimal predictor. (10 min)
13 Overview Lecture 2C. MLR - the basic model. (6 min)
13 Overview Lecture 2D. Biased regression methods: Ridge, Lasso, PLS, PCR. (14 min)
13 Overview Lecture 2E. All the methods are shrinking the MLR solution. (8 min)
13 Overview Lecture 2F. The overall modelling approach: Cross validation etc. (8.5 min)