Discovery Indonesian Medical Question-Answering Pairs Pattern With Question Generation

Penulis: Wiwin Suwarningsih, Iping Supriana, Ayu Purwarianti
Question generating refers to the task of automatically generating questions from various inputs such as raw text, database, or semantic representation. This paper concerns with the development of the novel question generating methods for Indonesian medical question answering. It begins by presenting an overview of the main methods and followed by showing the architecture of the proposed system in detail. Furthermore, this paper proposed a novel template-based approach to question a generation by combining a number of semantic roles through a method of generating both general and domain-specific questions. It was more focused on Question Generating with a discourse task involving the following three steps: (1) content selection of sentence (Problem, Intervention, Comparison and Outcome (PICO) categories), (2) question type identification, and (3) question answering construction. It is argued then that many ways in which the sentences of the answer to a question can be formulated and acquired from resources containing a large number of semantically related sentences.This approach achieved F-measure values of 0.88, precision 0,84 and recall 0,86 for active sentence.Finally, it can be concluded that the pattern matching criteria of the training set and semantic role labelling based on PICO frame can be reproducible with minimal expert intervention.

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 10, Number 14 (2015) pp 34217-34223 Research India Publications.

ISSN / ISBN / IBSN : 0973-4562

No. Arsip : LIPI-15064