Research And Applaication Of Automatic Question Answering System

Now is an Internet era and also is an era of information explosion. The user’s retrieving needs are gradually changing. Existing search engines generally only provide keywords combination rule based queries. The returned results which are sorted sets of web pages that contain the keyword or query conditions can’t afford user’s demand. How to use natural language processing techniques in answering system is an effective way to improve the traditional search engine’s performance. QA system’s input statement is expressed by the natural language. By extracting information from questions of inquiries and parsing out the user’s query intent, the QA system can accurately locate the user needs and extract the answer from the documents. By returning the answer to the problem instead of the distribution of the document to the users, it is more to meet the needs of the user to retrieve than the existing search engines.The main content of this paper is to study the QA system and its’ applications. First, the overall design scheme of an automated answering system are proposed in this paper, and then the three core modules: question classification module, keyword expansion module and answer extraction module are detailed designed. For question classification module, this paper introduces the Chinese word segmentation model and classification model, then proposes a combination of Naive Bayes and SVM algorithm for the question classification. For the keyword expansion module, this paper introduces the keyword expansion process and presents an algorithm combination based on synonyms dictionary and HowNet. For answer extraction module, this paper proposes an answer extraction algorithm based on semantic dependencies. At last, this paper presents an automated answering system application examples. Based on Nginx module development, this paper extends the original instruction set of Nginx and then packages the QA processing module to the dynamically service object. On this basis, this paper develops an automatic question answering system server which can be used in a real production environment.