| In recent years,with the continuous development of Internet technology,online education has become an important way of learning.With the injection of a large number of online learners,the demand for teaching resources is growing with each passing day,and the scale of digital education resources is also expanding.In the context of advocating smart education and personalized education,how to effectively organize,aggregate and evaluate the massive crowdfunding and innovation education resources is the primary prerequisite for the construction and management of a digital education resource platform.Under the crowdfunding and innovation model,it is difficult to deal with a lot of digital education resources,and the types of resources are diversified,and it consumes a lot of costs in manual classification and retrieval.When students learn online,they need to sort out the concepts in the curriculum resources,and the concepts of crowdfunding and innovation education curriculum resources are often hidden in the content,which requires learners to self-summarize and organize,which increases the cost of learning.Therefore,how to use big data and artificial intelligence methods to construct a knowledge map of crowdfunding and innovation education resources is one of the difficulties in the construction of digital education platforms and resources.At the same time,crowdfunding and innovation education resources are independently created and uploaded by users.With the rapid growth of resource contributors,the quality of resources has become uneven.Therefore,in the process of generating the knowledge graph,it is necessary to evaluate the quality of the generated candidate resources and eliminate unqualified resources to ensure the quality of the knowledge graph.This thesis studies the generation and quality evaluation of educational resource knowledge graphs under the crowdfunding and innovation model.The main work is divided into the following three parts:(1)Aiming at the problem of generating knowledge graphs of educational resources,this thesis proposes an improved named entity recognition and relation extraction method,which combines traditional knowledge extraction methods with the extraction of curriculum concepts in educational resources,and combines the BERT migration learning model With the custom encoder,it effectively solves a series of problems in the extraction of educational terms,and extracts the upper and lower relations of concepts based on prior knowledge such as types,and finally conducts experiments based on the educational subject data set.(2)Aiming at the quality of educational resources under the crowdfunding and innovation model,this article takes encyclopedia entries as the research object,and proposes a method for evaluating the quality of candidate resources based on the knowledge graph generation of a multi-decision model,using a method of multi-classifier ensemble learning,to score and rank candidate resources,and conduct experiments based on the Baidu Encyclopedia data set.(3)On the basis of the above methods,this thesis builds a set of educational resource knowledge map generation and evaluation system,which integrates functional modules such as knowledge extraction,relational reasoning,triple quality evaluation,etc.,to provide an educational resource knowledge map for the digital education platform The generated full link capabilities are displayed in a visual manner.This thesis conducts research from two aspects of theory and application,provides a solution for the generation and quality evaluation of the knowledge graph of crowdfunding and innovation education resources,and promotes the development of digital education. |
