Design and Optimization of an Intelligent Question Answering System Integrates Natural Language Processing and Knowledge Graph

With the rapid development of information technology, intelligent question answering system (QAS) has become an important tool for users to obtain information. This paper aims to design an intelligent QAS which integrates naturallanguage processing (NLP) and knowledge graph (KG), and optimize its performance. By effectively combining NLP with KG technology, the system can accurately understand the semantic information of users’ questions and provide users with accurate and comprehensive answers with the support of KG. Firstly, this paper introduces the development background and technical challenges of intelligent QAS, especially the important role of NLP and KG in the system. Thenthe architecture design of the system is described in detail, including input layer, processing layer and output layer, and the main functions and technical realization of each layer are introduced. In the aspect of system optimization, this paper puts forward query pruning strategy based on heuristic rules, caching mechanism and semantic understanding ability improvement strategy based on BERT (Bidirectional Encoder Representations from Transformers) model, anddesigns a set of regular updating and maintenance mechanism of KG. Through the experimental verification of two public data sets, WikiQA and TREC QA, the system shows high accuracy and recall rate, which proves the reliability and effectiveness of the intelligent QAS designed in this paper in practical application. This research not only enriches the theoretical basis of intelligent QAS, but also provides new ideas and methods for practical application in related fields.