The paper concerns the measurement and forecasting of technological change, a topic relevant to many high-tech organizations and their customers. We revisit recent and classic data sets from technology forecasting data envelopment analysis (TFDEA) research and technometrics in light of a new visualization technique known as t-Distributed Stochastic Neighbor Embedding (t-SNE). The technique is a non-linear visualization technique for preserving local structure in high-dimensional spaces of data. The technique may be classified as a form of topological data analysis. Specifically each point in the space represents a potential technological design or implementation, and each line segment in the space represents a local measure of technological improvement or degradation. We hypothesize six distinct kinds of performance development in technology within this space including the frontier, the fold, the salient, the soliton, and the lock-in. Then we examine the spaces to determine which kinds of development are the best explanations for observed development. The technique is not extrapolative, and therefore cannot supplant existing technometric methods. Nonetheless the approach offers a useful diagnostic to existing technometric methods, and may help advance theories of technological development.
|Title of host publication||Proceedings of PICMET '14 Conference|
|Subtitle of host publication||Portland International Center for Management of Engineering and Technology; Infrastructure and Service Integration|
|Publication status||Published - 13 Oct 2014|
- technological change
- visualization approach
- data analysis