| Intangible cultural heritage is the precious resource of human civilization,history and culture.At present,the majority of Chinese intangible cultural heritage are managed and protected through the provincial and municipal intangible cultural heritage websites and intangible cultural heritage databases in a few fields.And its knowledge expression is mainly in the field of related literature,databases,common domain of intangible cultural heritage ontology.Such expression in the form of websites and databases has some problems,such as weak knowledge relevance,heterogeneous multi-sources and incomplete knowledge records.Therefore,using modern computer technology to protect,manage and utilize intangible cultural heritage is the direction for the scholars to study and research.In view of the characteristics of multi-source heterogeneity and knowledge dispersion of intangible cultural heritage knowledge resources,this study proposed a method using ontology technology and text knowledge extraction technology to construct knowledge ontology in the field of intangible cultural heritage architecture,including the whole process of design,modeling,storage and release of knowledge ontology.Using ontology technology to realize the digital management of intangible cultural heritage resources is conducive to the protection,utilization and sharing of intangible cultural heritage resources.Deep learning algorithm is applied to realize knowledge extraction of named entities in intangible cultural heritage unstructured text,and enrich knowledge of domain ontology.The main research is as follows:(1)Construct ontology corpus of knowledge in the field of intangible cultural heritage architecture.In this study,structured data such as CN-DBpedia and Ownthink,semi-structured data such as tables and lists from China Intangible Cultural Heritage Website,national and provincial intangible cultural heritage websites,and unstructured data such as related academic literature and Baidu encyclopedia entries were collected through crawler technology.They are used as a corpus for the construction of knowledge ontology.(2)Design and modeling of knowledge ontology in the field of intangible architectural heritage.In the studies on ontology of many fields,the vocabulary of architecture field is highly professional,and no standard domain knowledge ontology has been formed.Meanwhile,few people know about the architecture field of intangible cultural heritage.Therefore,this study applied CIDOC CRM,an ontology framework in the field of cultural heritage protection,to the construction techniques of China traditional wooden structures.Combined with the characteristics of the field of architecture,the structure of CIDOC CRM model was selected and expanded to define the conceptual categories and attributes of the ontology of intangible cultural heritage knowledge in the field of architecture.The initial instantiation was carried out by Protégé ontology editing tool.(3)Instance expansion of the knowledge ontology in the field of intangible heritage architecture.In order to obtain more abundant ontology instances and fine-grained domain knowledge,this study defined the named entity types,annotated the collected unstructured texts with BIO,and took self-built annotated corpus as the pre-training data set.By comparing the training effects of BERT-WWN and Word2 vec model,BiLSTM and LSTM model,and CRF model and softmax multi-classifiers,the BERT-WWNBiLSTM-CRF model is finally selected to realize the model training of self-defined named entities and complete the named entity recognition task.The acquired named entities are batch-imported as instances through the Cellfie module built in Protégésoftware to complete the expansion of ontology instances and enrich the content of knowledge ontology.(4)Storage and release of knowledge ontology in the field of intangible architectural heritage.The constructed knowledge ontology is stored in the Neo4 j graph database through data mapping,and visualized in the form of knowledge graph.In order to realize ontology knowledge releasing,this study designed a knowledge graph application platform based on Chinese traditional wooden architecture construction techniques based on Django framework and Echarts visualization tool,intuitively displaying the acquired knowledge in the field of intangible architectural heritage and releasing fine-grained knowledge for the resource in the field of intangible architectural heritage.The application platform realizes knowledge management and semantic association of 19 intangible cultural heritage projects under its list.The designed application platform can realize knowledge graph query based on intangible cultural heritage project entities and entity relationships,and build a bridge between users and digital resources.The results show that it is feasible to construct ontology model of traditional architectural intangible heritage knowledge by using the above methods.The created text effectively connects the various fragmented knowledge.This method can be used for knowledge management in the field of ancient architecture,and will play an important role in the digital protection and rational utilization of intangible cultural heritage in the field of architecture.It is significant for the establishment and sharing of structured data resources of intangible cultural heritage projects such as traditional dance,drama,folk literature and folk customs.Figure [32] table [19] reference [75]... |
PDF Full Text Request