Research And Implementation Of Intelligent Summarization Algorithm Of QA Based Mass Data For Chronic Diseases

The vigorous development of the Internet has brought a lot of data accumulation. The health care industry as a kind of very important data source that has been attracted more and more attentions.Today, online medical service products have sprung up, including online question-and-answer service, with its convenient and effective, reply in a timely manner, strong professional characteristics are recognized and received by more and more people. Therefore, the research on the China currently active higher health portal, we found that there has produced hundreds of millions question and answer data after so many years accumulation. Among them, diabetes, as the representative of the chronic disease which asked the most warmly, how to make full use of the existing data for people who has the similar problems in the future is the key to the intelligent question answering system.The intelligent question answering system studied in this paper is based on the historical data to analyze and process, so as to provide a timely and effective solution for the new problem. Although there has been a part of the intelligent question answering systems abroad, but it has some limitations, such as limit the type of the problems. And there are a lot of new problems to be solved in the process of Chinese data.The main contributions of this paper are as follows: first, after a careful analysis of the structure of the question and the answer, we extracted the entities, which contain some very important information, through the entity we can clearly express the answer semantic; and then combined with the answers' TFIDF value, the question keywords, sentence features, and the question similarity to score the answer. The experimental results show that the algorithm added with the entities improves the quality of answer summary. Second, this paper also realized the optimization of the interface, on the one hand, we research the typical problems of chronic diseases, the users can obtain the corresponding answers directly click on these problems, so as to improve the efficiency; on the other hand, We also improved the presentation of the answer for the health consumer. The single answer is to display the entire section of the original answer data, The list of fragments are fragment that contain a wealth of information after paragraph segmentation, The combination of fragments is a brief, high accuracy passage by assembling higher score fragments.Finally, we measured the performance of the entity extraction, and the overall evaluation of this system, the results show that the system has better applicability compared to other health services.