ABSTRACT
Throughout scientific research, measured time series are the basis for characterizing an observed system and for predicting its future behavior. A number of new techniques (such as state-space reconstruction and neural networks) promise insights that traditional approaches to these very old problems cannot provide. In practice, however, the application of such new techniques has been hampered by the unreliability of their results and by the difficulty of relating their performance to those of mature algorithms. This chapter reports on a competition run through the Santa Fe Institute in which participants from a range of relevant disciplines applied a variety of time series analysis tools to a small group of common data sets in order to help make meaningful comparisons among their approaches. The design and the results of this competition are described, and the historical and theoretical backgrounds necessary to understand the successful entries are reviewed.
