Data mining on chemical process data (consulting)

Technologies used: Matlab and toolboxes. 
Developed in 2005.

I worked with Prof. Thorsteinn Rognvaldsson from Halmstad University on a consulting project for Eka Chemicals from Sweden.

It involved the multi-variate data mining with several variables representing measurements from the chemical process. The main goal was to find out about the important variables influencing negatively the process.

For that, machine learning tools such as logistic regression, multi-layer perceptrons, and support vector machines were used (in Matlab).