Department Process Engineering

Software: SCS

The SCS toolbox allows to fit shape constrained spline functions based on second order cone programming (SOCP). In conjunction with the SPIKE_O toolbox, this toolbox also allows to identify the optimal transitions between qualitative primitives, thanks to proven bounding procedures. It has been tested in Matlab (R2010a, R2012b, R2014b) and it is publicly available under the GPLv3 license. Support  for Octave is not available yet (Octave-to-MOSEK Optimization Interface remains to be tested).

SCS v1.0. This version allows fitting of SCS functions, however restricted to univariate functions without discontinuities. Optimization can be executed via CVX or Mosek. The SCS toolbox makes use of an extended version of the FDA toolbox which includes natural spline functions. The following version of this toolbox includes all necessary functions:

SCS v2.x. These versions allow fitting of univariate SCS functions with and without discontinuous trends. Optimization can be executed via Mosek. There is no CVX-based implementation. Apart from the Mosek software, these versions are self-contained.

Citing the SCS toolbox

Are you using the SCS toolbox in research work to be published? If so, please include explicit mention of this toolbox in your publication. We suggest language such as this: "To find these qualitative representations we used the SCS toolbox, a package for spline-based qualitative trend analysis [1]." with one or more of the following entries in your bibliography:

Villez, K., Venkatasubramanian, V., Rengaswamy, R. (2013). Generalized Shape Constrained Spline Fitting for Qualitative Analysis of Trends. Computers and Chemical Engineering, 58, 116-134.

Habermacher, J., Villez, K. (2016). Shape Anomaly Detection for Process Monitoring of a Sequencing Batch Reactor. Computers and Chemical Engineering, 91, 365-379.