GOOD vs BAD; What´s the difference

Finding out the difference between groups of samples (observations, time points, items, cases, runs, individuals, …) is one of the most central questions in contemporary data analytics and data mining investigations. What´s the difference between good or bad manufacturing conditions? Or between real or fake product? Or between wildtype or genetically modified species? Or between single malt or blended whisky?

Attend this webinar and discover the analytical powers of OPLS-DA. Whenever possible it is strongly recommended to formulate the discriminant problem as a two-class problem. This is because the resulting OPLS-DA model can then only have one predictive component and this is the easiest situation for revealing the best discriminating variables. With OPLS-DA comes a number of tailor-made plots assisting in the identification of such discriminating variables. These plots, denoted the S-plot, S-line plot, and SUS-plot, are exemplified.

Topics for this webinar:

  • Basic principles for the two-group discriminatory problem
  • Main outcomes of OPLS-DA
  • Score plot to visualize the difference
  • Prediction of new observations, samples
  • Loading plot to define variables that define the difference
  • 2D and 3D variable plots of the important variables
  • S-plot, S-line plot, SUS-plot