Font Size: a A A

Research On Chinese Intelligent Question Answering Model Based On Speciality-Hownet

Posted on:2007-12-04Degree:MasterType:Thesis
Country:ChinaCandidate:Z H LiFull Text:PDF
GTID:2178360185958900Subject:Computer applications
Abstract/Summary:PDF Full Text Request
In recent years there appear in large numbers of network-based instruction platforms.Because of the separating between teachers and students in space and time in network-based instruction,the answer for questions based on net,as a key way to communicate between students and teachers.become more and more important.At present,there are four answering ways e-mail,discussing on-line,query in key word and FAQ in network-based instruction platforms.Each of them has its own shortcoming:(1) in e-mail,teacher not only can not answer question in time but also answer the same question for different students;(2)in discussing on-line,teacher must be on-line all the day;(3)in query in key word,student must have the capacity of extracting key word;(4)in FAQ ,only the figure of key word,not the semantic of it,is concerned with.Aiming to solve the existing problems of Chinese Q/A system,this paper is going to state,by search,that the comprehension of sematic and semantic similarity computation for students question is the key of Chinese Q/A system ;that blur match based on sematic comprehension and similarity computation is an applicable way to improve the precision of answer in Chinese Q/A system.Based on the above-mentioned idea,according to the hownet,this paper provides the definition of the speciality-sememe and the way to extract it,the definition of speciality-semsms tree and the way to build it ,the rule and way to mark speciality-key-words,and introduces the realization of the semantic similarity computation of speciality-sememes and speciality-key-words by improving semantic similarity computation of hownet-sememes and hownet-words,then,based on all of the above-mentioned. As a whole,speciality-hownet is built, by transfering speciality-hownet and sentence similarity computation into Chinese Q/A system,we proposes a CIQAMS(Chinese Intelligent Question Answer Model Based on Speciality-hownet),we have designed a speciality-hownet model to realize the understanding of semantic and similarity computation for speciality-key-words.Guided by the designing ideas of CIQAMS,we have designed a CMHSCCES(Chinese Modern History Sematic Comprehension and Computation Experiment System) to verify the functionality of the sematic comprehension and computation.And the experiment has proved that a better testing result is gained by adding the comprehension of sematic relation and sematic computation.
Keywords/Search Tags:Chinese intelligent Q/A System, Speciality-hownet Model, Understanding of Sematic, Sematic Similarity Computation, Natural Language
PDF Full Text Request
Related items