首页|基于词向量和卷积神经网络的税务问答系统

基于词向量和卷积神经网络的税务问答系统

扫码查看
税务咨询问答业务量大,要求专业程度高,且近年来新出台的税收优惠政策较多,纳税人针对税务相关法律法规的咨询需求也是日益见长.现有的税务咨询方式大都是办税大厅现场咨询或电话咨询,税务工作者需要针对纳税人的问题,不断学习和查阅相关政策法规,使得税务机关面临着巨大的咨询压力.现如今虽然一些税务机关开始尝试利用互联网进行线上税务咨询,但大都是留言咨询,需要人工对问题进行解答,大量热点问题集中造成系统拥堵,且不具备即时性,不能有效解决当前问题.智能问答技术能够有效解决上述问题,基于该技术构建的税务问答系统具有效率高、实用性强等优势,对纳税人和税务工作者都提供了更多的便利.
Tax Q&A System Based on Word Vector and Convolutional Neural Network
Tax consultation Q&A business volume is large,requires a high degree of professionalism,and in recent years,the new tax preferential policies are more and more,taxpayers'demand for tax related laws and regulations is also growing.Most of the existing ways of tax consultation are on-site consultation or telephone consultation in the tax hall.Tax workers need to constantly study and consult relevant policies and regulations for the problems of taxpayers,which makes the tax authorities face huge pressure of consultation.Nowadays,although some tax authorities are trying to use the Internet for online tax consultation,most of them are message consultation,which requires manual answers to the problems.A large number of hot issues are concentrated,causing sys-tem congestion,and are not instant,which can not effectively solve the current problems.Intelligent question answering technology can effectively solve the above problems.The tax question answering system based on this technology has the advantages of high effi-ciency and strong practicability,which provides more convenience for taxpayers and tax workers.

tax consultationintelligent question and answernatural language processing comprehensionneural network

杜航、牟莉

展开 >

西安工程大学计算机科学学院 西安 710600

税务咨询 智能问答 自然语言理解 神经网络

2024

计算机与数字工程
中国船舶重工集团公司第七0九研究所

计算机与数字工程

CSTPCD
影响因子:0.355
ISSN:1672-9722
年,卷(期):2024.52(1)
  • 15