Chinese Named Entity Recognition Based on Knowledge Based Question Answering System

The KBQA (Knowledge-Based Question Answering) system is an essential part of the smart customer service system. a type QA (Question based on KB (Knowledge Base). It aims to automatically answer natural language questions by retrieving structured data stored in knowledge base. Generally, when receives user’s query, it first needs recognize topic entities such as name, location, organization, etc. This process NER (Named Entity Recognition). In this paper, we use Bidirectional Long Short-Term Memory-Conditional Random Field (Bi-LSTM-CRF) model and introduce SoftLexicon method for Chinese task. At same time, according analysis characteristics application scenario, propose fuzzy matching module combination multiple methods. can efficiently modify error recognition results, which further improve performance entity recognition. We combine into To explore availability some specific fields, power grid field, utilize grid-related original collected Hebei Electric Power Company our field. innovatively make dataset high-frequency word lexicon makes proposed perform better recognizing field grid. used cross-validation validation. experimental results show that F1-score improved reaches 92.43%. After processing using module, about 99% test set be correctly recognized. proves achieve excellent scenario work will also fill gap research intelligent customer-service-related technologies China.