ABSTRACT
Optimization is shown to be context-dependent and, thus, a functional setting of a text analytical system. We consider optimization in light of its input, its operations, and its output. Input can be readily changed through different levels of compression, or summarization, of the content. Operations to be considered for optimization include those that vary by time, location, language, and adjacent uses and applications. Outputs are the impact that the optimized system has on its environment – its users, its downstream algorithms, and various measurements of its overall acceptability. Thus, optimization is shown to be a process, rather than an end point, and as such is shown to be a functional text analytic. This alternative view to optimization is applied to important functional text analytical systems.
