Big data analytics often relates to addressing terabytes of data and using parallelization approaches to do what might
be considered basic calculations. Bigish data analytics (our word) relates to using much larger data than was previously considered
for relatively powerful modeling and optimization activities.
Currently, related research involves speeding up and otherwise improve sequential Kriging optimization and subject matter
expert refined topic (SMERT) models. Both methods relate to human-in-the-loop "machine learning". In particular,
SMERT models relate to view that (some) experts can be trusted to influence models in desirable ways.
White Paper on Bigish Analytics
Discrete Event and FEM Simulation Optimization
Computer simulation is a well used practice in engineering. We
develop models that sometimes look very realistic visually with virtual people and machines making things or accomplishing
tasks (like voting).
When the model is made and it has
some degree of trustworthiness with validation, it makes sense to try out a lot of "what if" studies. Design of
experiments and simulation optimization are names of areas in which methods are generated for efficient what if experimentation
and recommended setting generation. In some cases, they offer the guarantee that all possible what ifs within a set have effectively
Some of my students and I developed methods building
on work by Schonlau and others which offer pretty efficient way to find desirable settings without too much experimentation.
Sequential Kriging Optimization
This data was cited in relation to the following publications.
Allen, T. T., Xiong, H., and Afful-Dadzie,
A. (2016). A Directed Topic Model Applied to Call Center Improvement. Applied Stochastic Models in Business and Industry,
Allen, T. T. and H. Xiong (2012). Pareto charting using multifield freestyle text data applied to Toyota
Camry user reviews. Applied Stochastic Models in Business and Industry, 28 (2), 152-163.
click here to download call center data
click here to download Camry report data
click here to download Iraq analysis requiring directed modeling
This data contains a collection from multiple sources. If you use it, please cite all of the following:
C. and Allen, T. T. (2003). An Alternative Desirability Function for Achieving ‘Six Sigma’ Quality. Quality and
Reliability Engineering International, 19, 227-240
Allen, T. T., R. W. Richardson, D. Tagliabue, and G. Maul (2002).
Statistical Process Design for Robotic GMA Welding of Sheet Metal. The Welding Journal, 81, 5, 69s-77s
Allen, T. T.,
W. Ittiwattana, R. W. Richardson, and G. Maul (2001). A Method for Robust Process Design Based on Direct Minimization of Expected
Loss Applied to Arc Welding. The Journal of Manufacturing Systems, 20, 5, 329-348 (http://www-iwse.eng.ohio-state.edu/ISEFaculty/allen/AllenIttiwattanaRichardsonMaul2001.pdf.
click here to download large arc welding dataset