[1]陈 强,代仕娅.基于行业知识与知识图谱的会计案防智能平台[J].计算机技术与发展,2022,32(01):180-185.[doi:10. 3969 / j. issn. 1673-629X. 2022. 01. 031]
　CHEN Qiang,DAI Shi-ya.Intelligent Accounting Risk Management Platform Based on IndustrialKnowledge and Financial Knowledge Graph[J].,2022,32(01):180-185.[doi:10. 3969 / j. issn. 1673-629X. 2022. 01. 031]
点击复制

基于行业知识与知识图谱的会计案防智能平台()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年01期
页码:
180-185
栏目:
应用前沿与综合
出版日期:
2022-01-10

文章信息/Info

Title:
Intelligent Accounting Risk Management Platform Based on IndustrialKnowledge and Financial Knowledge Graph
文章编号:
1673-629X(2022)01-0180-06
作者:
陈 强1 ; 代仕娅2
1. 兴业银行 信息科技部,上海 201201;
2. 兴业数字金融服务股份有限公司,上海 201210
Author(s):
CHEN Qiang1 ; DAI Shi-ya2
1. Department of Information and Technology,Industrial Bank Co. ,Ltd. ,Shanghai 201201,China;
2. CIB Digital Financial Services Co. ,Ltd. ,Shanghai 201210,China
关键词:
会计案防; 金融知识图谱; 智能化平台; 图分析; 关联特征
Keywords:
accounting risk management; financial knowledge graph; intelligent platform; graph analysis; connected features
分类号:
TP39
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 01. 031
摘要:
商业银行会计风险案件日益频发,传统人工核查的方法难以及时有效地核查风险,为此提出了一种将会计案防领域行业知识与大数据、知识图谱等技术相结合的智能化平台建设方案。 平台采用大数据技术进行会计基础业务指标的加工计算,依托图数据库构建账户资金交易知识图谱,采用图分析技术完成账户关联风险特征的挖掘提取;最后将基础业务指标与图关联特征相结合,构建了完整的会计案件风险识别规则体系,实现浅层经验知识与深层挖掘知识的融合应用。以银行活跃个人账户作为实验样本,实验结果表明,该智能化平台能实现规则配置的全流程、自动化线上运行,能更高效准确地识别出可疑风险账户,与传统核查方式相比极大提高了风险预警的及时性、精准性与前瞻性。
Abstract:
The frequency and complexity of commercial banks’ accounting risk events increase rapidly in recent years,which brings aboutgreat challenges to traditional risk verification approaches. Aimed at preventing accounting risk more effectively, a new intelligentplatform solution which combines accounting industry knowledge, big data technology, and financial knowledge graph is proposed.Firstly,various big data related technologies are fused to ensure capacity for heterogeneous data storage,feature engineering,and graphcomputing. Secondly,massive graph connected features are extracted from the financial knowledge graph of transaction accounts throughgraph algorithms to deeply reveal the hidden risks. Finally,the industry knowledge-based features and the graph connected features aremerged to form an abundant and flexible rules system for accounting risk recognition. Taking bank active individual account asexperimental sample,it is showed that the intelligent platform makes the rules configured automatically during whole process and alerts ac-counting risks in more timely, accurate and dynamic ways compared to traditional human verification method, demonstrating that theplatform can significantly enhance the digitalization and automation level of the accounting risk management.
更新日期/Last Update: 2022-01-10