基于自然语言处理的铁路客运营销分析智能对话系统研究
Research on intelligent dialogue system for railway passenger transport marketing analysis based on natural language processing
李仕旺 1江琳 2王桂林3
作者信息
- 1. 中铁程科技有限责任公司,北京 100081
- 2. 中国铁道科学研究院集团有限公司 电子计算技术研究所,北京 100081
- 3. 北京经纬信息技术有限公司,北京 100081
- 折叠
摘要
为提高铁路客运营销数据分析能力,研究开发了铁路客运营销分析智能对话系统,为铁路客运营销业务人员提供一种基于人机对话的数据分析工具.该系统包括语音识别、自然语言文本处理、智能数据挖掘、智能应答4个主要功能模块;利用语音唤醒和语音识别技术采集语音数据,通过神经网络模型将语音数据转换成自然语言文本;建立自然语言文本预处理模型,完成基于规则的词法句法分析方法,使用长短期记忆神经网络实现语义理解,确定用户意图;基于Bert模型的Text-to-SQL技术,将自然语言文本数据转换成数据查询SQL语句,构建智能Agent完成数据挖掘分析,生成数据分析结果;最后,运用语音合成技术和数据可视化技术,将数据分析结果转换为用户应答信息.
Abstract
To improve the data analysis capability of railway passenger transportation marketing,an intelligent dialogue system for railway passenger transportation marketing analysis has been developed,providing a data analysis tool based on human-machine dialogue for railway passenger transportation marketing business personnel.The system includes four main functional modules:speech recognition,natural language text processing,intelligent data mining,and intelligent response.It uses voice wake-up and speech recognition technology to aquire voice data,and converts the voice data into natural language text through neural network models.A natural language text preprocessing model is established to complete rule-based lexical and syntactic analysis methods.Then,long short-term memory neural networks is used to achieve semantic understanding and determine user intent.Bert-based Text-to-SQL model is employed to converts natural language text data into data query SQL statements and intelligent agents are constructed to complete data mining and analysis,and generates analysis results.Finally,speech synthesis and data visualization are used to convert the analysis results into reply to user.
关键词
客运营销分析/智能对话系统/自然语言处理/语音识别/深度学习/语义理解/数据挖掘/语音合成/数据可视化Key words
passenger transport marketing analysis/intelligent dialogue system/natural language processing/speech recognition/deep learning/semantic understanding/data mining/speech synthesis/data visualization引用本文复制引用
基金项目
北京经纬信息技术有限公司基金项目(DZYF23-09)
出版年
2024
NETL
NSTL
万方数据