Text Analysis Pipelines by Henning Wachsmuth

Author:Henning Wachsmuth
Language: eng
Format: epub
Publisher: Springer International Publishing, Cham

Input Dependency of Optimized Scheduling. Lastly, we investigate in how far the given input texts influence the quality of the pipeline constructed through optimized scheduling . In particular, we evaluate all possible combinations of using the Revenue corpus and the CoNLL-2003 dataset as input for the three main involved steps: (1) Determining the run-time estimations of the pipeline’s algorithms, (2) scheduling the algorithms, and (3) executing the scheduled pipeline. To see the impact of the input, we address only the query , where our k-best A search approach is most successful, i.e., . We compare the approach to the greedy baseline , which also involves step (1) and (3). This time, we leave both k and the number of training texts fixed, setting each of them to 20.Table 4.3The average execution times in seconds with standard deviations of addressing the query using the pipelines scheduled by our 20-best A search approach and the greedy baseline depending on the corpora on which (1) the algorithms’ run-time estimations are determined, (2) scheduling is performed (in case of 20-best A search ), and (3) the pipeline is executed.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.