ETA Research

It is imperative to pick the right tool to solve a particular problem. ETA researchers frequently use the following approaches.

Technology Forecasting

Yogi Berra supposedly said "It's difficult to make predictions, especially about the future." Unfortunately that's true - especially in the rapidly moving areas of high technology. Companies may need to set targets or design envelopes for new product development efforts years in advance but how do you do this? This requires forecasts of where a technology will be years down the road. The Extreme Technology Analytics group at Portland State University has developed a new approach called Technology Forecasting Using Data Envelopment Analysis or TFDEA for short. TFDEA allows for a broad characterization of the technology without requiring explicit or fixed tradeoffs between these items. For example, an early paper extended Moore's Law to allow for manufacturing, design, and performance factors simultaneously and better predicted the competitive microprocessors in the 1990s.

Since TFDEA was first developed in the early 2000s, applications have included a wide range of industries including fighter jets, telecommunications protocols, commercial passenger aircraft, hybrid electric vehicles, and many more. International researchers have used it in the Netherlands, Turkey, Korea and China.

Quantitative Benchmarking

Dr. Anderson created possibly the first web page for Data Envelopment Analysis back in 1993 using vi, long before the availability of friendly GUI HTML editors. Since that time, he has continued to stay active in the field of DEA.


In this resource constrained era, we need to make the best or optimal use of our limited resources. Optimization can be done at the design level such as the layout of a chip or the parameters of a mechanical design. Our team applies it to a variety of areas such as staff scheduling for hospitals, room assignments for Naval training, and test vectors for semiconductor chips. A novel and interesting one was to examine the issue of realigning high school sports conferences around the state of Oregon to minimize travel time and cost.

Data Mining

Many organizations are buried in data but have a hard time leveraging it for real impact. In this case, data can be like a rich vein of minerals underground. Data mining helps us to look at the data in new ways to mine it for useful information. A major local company was sitting on years of lab reports collected over time and issued the team a data mining challenge - tell them something their years of experience, Master's and PhD researchers did not know. A minimal explanation of the data was given so as to not bias the miners by industry preconceptions. The challenge was so successful, it led to multiple follow-up funded projects. It is important to have a variety of tools in your toolbox. We've used a wide range of approaches ranging from neural networks and data envelopment analysis to the more common multivariate methods such as multiple logistic regression and cluster analysis.