Named Entity Recognition in Educational Psychology Based on the Bert Model

The abundance of resources in the network environment leads to a large amount of information in educational technology, low cognitive efficiency of learners, inability to focus their attention, and ultimately deviation from the learning objectives and inability to complete specific learning tasks. In order to solve these problems encountered by learners in e-learning, a combined BERT-BiLSTM-CRF named entity recognition method for educational technology main courses is proposed with the aim of improving learners' learning efficiency. First, we construct a dataset for named entity recognition in the main course of educational technology, and use the BERT language model for text feature extraction to obtain the word granularity vector matrix; then we use the bidirectional long short-term memory (BiLSTM) to extract the relationship between the input utterance and the context of the word; and we use the conditional random field (CRF) to extract the global optimum based on the dependency relationship between the labels. The global optimal output tag sequence is extracted based on the dependencies between the tags; and the named entities of the main courses of educational technology are obtained.