基于商品评论语义分析的情感知识图谱构建与查询应用  

Construction and Query Application of Emotion Knowledge Graph Based on Semantic Analysis of Product Reviews

作　　者：由丽萍[1] 郎宇翔 

机构地区：[1]山西大学经济与管理学院,山西太原030006

出　　处：《情报理论与实践》2018年第8期132-136,131,共6页Information Studies:Theory & Application

基　　金：国家自然科学基金项目"基于FrameNet的中文评价词汇本体构建与观点挖掘研究"的成果之一;项目编号:71403154

摘　　要：[目的/意义]知识图谱具有强大的语义理解能力和开放互联能力,建立中文情感知识图谱可以实现智能化的情感信息检索,从而协助用户优化其购买决策,也有助于产品和服务质量的改进。[方法/过程]基于框架语义理论建立细粒度的情感分类词典和语义角色分析模型,通过在线评论的情感语义标注构建情感知识图谱,在此基础上应用本体查询工具,实现情感知识的个性化、智能化查询。[结果/结论]以手机商品评论为实验数据,实现了基于概念的综合查询和语义推理,展示了情感知识图谱在智能化检索服务中的应用优势。[Purpose/significance] Knowledge graph has powerful semantic understanding capability and open interconnect capability,so establishing a Chinese emotion knowledge graph can help to achieve intelligent emotional information retrieval,which assists users to optimize their purchasing decisions and helps to improve products and services quality. [Method/process] This paper establishes a fine-grained emotion classification dictionary and a semantic role analysis model based on the framework semantics theory. It constructs an emotion knowledge graph through emotional semantic annotation of online reviews and based on this,realizes a personalized,intelligent query of emotion knowledge with ontology query tools. [Result/conclusion] Based on the data of mobile phone products reviews,concept-based comprehensive query and semantic reasoning are realized,which demonstrates the application advantages of the emotion knowledge map in intelligent retrieval services.

关 键 词：框架 知识图谱 信息检索 语义分析 在线商品评论 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]