Construction Of Knowledge Map Of Ancient Poetry And Applied Research Driven By Knowledge Service

With the development of computer technology,and under the background of the country’s preparation of the "14th Five-Year Plan for Cultural Industry Development" to promote the high-quality development of the cultural industry,the introduction of deep learning and knowledge mapping technology into traditional Chinese literature is one of the current research hotspots.Ancient poetry is the treasure of Chinese traditional literature,which is concerned by scholars.Therefore,use deep learning and knowledge graph technology to process and study the vast amount of ancient poetry materials on the Internet has important academic significance and application value,and provides a way of thinking for broader and deeper research in the field of Chinese traditional literature.This thesis obtains user’s demand for ancient poetry knowledge service through questionnaire,constructs ancient poetry knowledge graph driven by knowledge service,and studies the application of knowledge service based on ancient poetry ontology model.The main contribution is to design the ancient poetry ontology model(APOM),construct the background corpus of ancient poetry creation,and put forward a classification model of ancient poetry theme based on MF-Text CNN-Bi LSTM,which combines multiple features(MF),Text CNN(Cable News Network)convolution layer and Bi LSTM(Bi-directional Long Short-Term Memory).The main contents are as follows.(1)Analyze the demand for knowledge service of ancient poetryThis thesis uses the method of questionnaire to investigate the demand of knowledge service in the field of ancient poetry,constructs the mapping structure of knowledge service of ancient poetry and obtains the knowledge set of ancient poetry according to the mapping relationship of "knowledge service demand-knowledge of ancient poetry".(2)Design APOM and construct knowledge graph of ancient poetryAPOM is designed according to the mapping relationship between "Ancient Poetry knowledge-Ancient Poetry Ontology Model".Use "BIO" format to mark the background data of ancient poetry creation to construct the background corpus of ancient poetry creation.The entity recognition model based on Bi LSTM-CRF is used to identify the time and place entities of the ancient poetry creation background corpus,and other models are used to conduct comparative experiments to verify the effectiveness of Bi LSTM-CRF on the ancient poetry creation background corpus.According to APOM and using Neo4 j to store entity-relation triples,the task of constructing knowledge graph of ancient poetry is completed.(3)Research on the application of knowledge service based on APOMFor theme classification of ancient poetry,this thesis first constructs the combination of image word collections of ancient poetry to put forward the marking rules of ancient poetry theme classification.And then puts forward the MF-Text CNN-Bi LSTM ancient poetry theme classification model to explore whether the title characteristics of ancient poetry have influence on the classification of ancient poetry theme.For the time and space trajectory representation of ancient poetry,we will explore the reasons for the change of poet’s creation themes from the perspective of the time and place of creation of ancient poetry.For visual knowledge retrieval,it provides personalized knowledge retrieval service for users to show the semantic relationship between ancient poetry knowledge.