
52 Text Mining and Visualization: Case Studies Using Open-Source Tools
raises the aggregate error to the first power (i.e., no change). This could be changed to find
the square root of the aggregate if desired. The operator is left in as a placeholder only.
The final part of the operator chain calculates a custom performance (operator 9). It
does by using the data value of the single aggregated error calculated in the preceding
operators. By this method, the sum of the errors for each example is reduced to a single
error value and this is the performance for the values of the macros A, B, and C when
calculating the Zipf-Mandelbrot fit. The last operator ...