Ontology Automation Of Sharing Economy Based On Deep Learing And Neural Network

Under the tide of the integration of everything in the world and the intelligent technology society,sharing economy has become a new hot spot of the current economy.However,as a new academic discipline,there are some pro-blems in the field of sharing economy,such as the confusion between theory and concept,and the ambiguity between concepts.Secondly,in real life,there is a lack of sufficient research on some factors that affect consumer behavior and manufacturer's profit and supply-demand relationship.Ontology is a rare and widely used knowledge representation method.Therefore,this paper proposes to build the ontology of sharing economy.The main work is as follows:(1)Corpus preprocessing and corpus building.By using the combination of focused crawling and incremental crawling to use the crawling queue,valuable shared economy corpus is screened out.We select the rough corpus,complete the corpus tagging and part of speech tagging,and build the ontology corpus in the field of sharing economy.(2)Construction of named entity recognition model in the field of sharing economy.On the basis of our shared economy corpus,we select word vector,character vector,context embedding vector and part of speech feature vector as the text embedding features of the named entity recognition model in the shared economy domain,and propose a named entity recognition model that integrates two attention mechanisms and multiple features.Our model goes beyond the baseline method in OOV problem and model performance.(3)(3)Construction of semantic relation extraction model in the field of sharing economy.In this paper,we extract the semantic relationship of named entities in the field of sharing economy,and select the word vector and position vector to form the sentence vector as the input of the relationship extraction model.We use an improved attention mechanism to fuse the input eigenvectors with the att-wp-blstm model.Using this improved attention mechanism IATT in the input layer,we can improve the performance and recognition rate of noisy sentences by ranking the vector features at the sentence level.Compared with the traditional clustering and association rule algorithm,the accuracy of the model is greatly improved.(4)Using protégé tool to build ontology of sharing economy.In this paper,we use the existing ontology tools to build the sharing economy ontology.We choose owl as our ontology language to import into the protégé tool to realize the visualization of sharing economy ontology.