A Pattern Matching Algorithm in P2P Mobile QA System Answer Extracting Process

Abstract

Mobile automatic question and answer (QA) system is more and more popular as a mobile value-added business. But the current mobile QA systems always have low query precision and recall ratio. In this paper, a new algorithm by pattern matching in p2p mobile QA system answer extracting process is designed. It expresses the document as triad and extracts the question and answer sentence from the triad based documents. Simulation experimental analysis results show that under ideal conditions the systempsilas average responding time is less than 58 ms and precision is above 66%, and the recall ratio is upper than 52% when there are no more than 3000 simultaneous query nodes, and so the completeness and feasibility of the systempsilas design is proved.