Research And Implementation Of Intelligent Question Answering System

In the reason of there are not enough time for the teachers and students to communicate andthe efficiency is low in the curriculum of the education process, and the backlog problem areincreasing in the students’ learning process, and traditional teaching methods are fully ported tothe network that make the teaching mode is not flexible, then affect the quality of the teaching, soThe an intelligent question answering system is researched and designed.During keyword search of traditional Q&A platform, there are problems that accuracy is notenough, the answer is colorful and slower query, and Cannot be accurately positioned, then thebased keyword weighting algorithm is proposed, the algorithm is in the basis of the originallycharacterized based on keywords, and the method based on the similarity of the Semantic distancemethod, under current Chinese characteristics, combined with the problem sentence segmentationmethod, weighted keywords, Weighted keyword of sentence problems, each keyword based on theco-occurrence frequency value judgment of weights, and then determine whether each wordco-occurrence frequency value within a certain range to stores the keyword library, can fast andeffective positioning the questions and answers,and the accuracy is Improved also.In order to improve the accuracy of the answer that improved a sentence similaritycalculation method. Traditional method of similarity based on the meaning of a word distancereflects the composition of each word of the sentence deep semantic information. Optimized onthe basis of these two methods that consider the similarity of the two aspects of distance fromkeyword characteristics and sentence meaning, make as a sentence of two-dimensional vector toobtain a sentence of two-dimensional vector-valued, through the comparison and analysis of thesetwo aspects,, to obtain the degree of similarity of the sentence, then such an improved method toobtain additional information, a more accurate description of the content of the sentence.There are the basics of one-to-one or one-to-many mode used for traditional Q&A platformsystem is the lack of complete information resource, the information slower response mechanismsuch as lack of a many-to-many mode of answering. First constructed a concept dictionary library,the formation of the mapping between the concept and the concept of set, and then be classifiedbased on the concept of the disciplines, the concept of automatic positioning function and conceptnetwork build. Second expertise, basic knowledge of the contents of the formation of theknowledge stored institutions Intelligent Answering Knowledge Base, integrated classification,information resource library that information resources are more abundant is construct. And thenmatch the fast output mechanism to achieve many-to-many intelligent question answering system.Design and implementation of an intelligent question answering system, focused on theactual testing and use of the "C#Programming" course teaching, the verify keyword weightedalgorithm and the sentence similarity algorithm improved really can effectively improve theaccuracy and speed up the search for the answer, Improve the accuracy of39%, Information largerresponse time shortened, and achieved good results.