【摘要】 目前国内各大型银行日交易量近千万笔,日平均交易额超过千亿元,产生大量结构化和非结构化的交易数据,这些数据难以有效地管理.与此同时,犯罪手段不断翻新,犯罪金额越来越大,给银行业带来的冲击越加明显.如何处理海量数据并及时地应对风险对实施风险监管系统提出了极大挑战.基于D-Ocean的金融风险监控系统依托D-Ocean系统将海量数据高效地管理起来,同时利用基于内容的文本相似性检索、基于颜色和图像特征的图像相似性检索、基于关键帧的视频相似性检索等技术充分挖掘非结构化数据在金融风险控制中的实际应用.更多还原

【Abstract】 For the moment,the daily trading volume of the domestic large banks has been near 10 million,and the average daily transaction volume exceeds one hundred billion yuan.They bring about a large amount of structured and unstructured transaction data which is difficult to manage efficiently. At the same time,criminal means are ever-changing,amount of loss is ever larger,and the impact to the banking industry is becoming increasingly obvious.How to deal with huge amounts of data and timely risk strategies present great challenges to the implementation of the risk of regulatory system. Financial risk monitoring system manages the vast amounts of unstructured data efficiently relying on the D-Ocean system,while taking advantage of the content-based text retrieval,image recognition, and video similarity search technology to fully tap financial risk control in actual applications.更多还原