| With the development of artificial intelligence technology,the intelligent question answering system has been widely used in our daily life.Intelligent question answering system can be divided into two categories according to the scope of answering questions:open domain intelligent question answering system and restricted domain question answering system.Open domain intelligent question answering systems have made great progress,but these research results can not be applied directly to restricted domain.On the one hand,question answering systems for restricted domain not only need to accurately identify professional vocabularies but also require a lot of domain knowledge as a support.The topic comes from a strategic project of a listed pharmaceutical company.The main purpose of the project is to help users solve some medical and disease problems,and provide users with convenient and intelligent medical knowledge services,thereby enhancing user stickiness and expanding the pharmaceutical market.At present,there is no mature medical intelligence question and answering product in the industry.Researchers in the academic community have proposed an intelligent Q&A program based on the knowledge base.However,the scheme has the following deficiencies:(1)The use of a general word segmentation scheme leads to inaccurate vocabulary identification in the medical domain;(2)There is no medical knowledge base construction scheme and there is a problem of insufficient medical knowledge;(3)It was built only for Chinese medicine question,and the content of the answer is the topological structure of the knowledge map,which deviated from the nature of question and answering.The paper proposes a design scheme for the intelligent question answering system in the medical domain,and implements it for the solution.The work of the article mainly includes the following parts:(1)Data acquisition.Designing web crawlers for multi-source heterogeneous data,collecting data from web sites,medical e-commerce websites,and medical Q&A websites.After identifying medical entity and extracting the relationships between different medical entities,we can contruct a medical knowledge base which solves the problem of insufficient knowledge in the medical domain;(2)Chinese word segmentation.The use of the Chinese word segmentation technology based on the medical dictionary solves the problem that the medical vocabulary cannot be accurately identified;(3)Word vector construction.The use of word vector construction techniques based on co-occurrence matrix and principal component analysis solves the problem of low computational efficiency caused by the high dimension of the word vector;(4)The generation of answers.The answer is generated using pattern matching and semantic similarity matching.The test results show that the intelligent question answering system for medical domain realized in the paper can correctly understand the user's question and make a professional and accurate answer.The system performance is great.Throughput rate and response time are within a reasonable range. |
