基于流数据频繁项挖掘的可疑金融交易识别研究

针对目前基于静态数据挖掘的可疑交易识别方法在处理该类交易数据时所面临的困难与局限性,结合可疑金融交易的特征,设计了基于流数据频繁项挖掘的可疑金融交易识别算法。该算法改进了有损计数法,利用实时保留的具有较高重复度的历史数据项解决了数据处理过程中的过度删除问题,实现了对频度列表中项的及时更新,并依据从数据流中识别出的频繁项来发现可疑金融交易线索。仿真实验结果验证了该算法的可行性和有效性。

Research on Suspicious Financial Transactions Recognition Based on Frequent Item Digging in Stream Data

Aiming to overcome the dificulty and limitation confronted by the method to recognize the suspicious transaction based on the digging of the static data when processing the data of that type,and combining the specific features of the suspicious transaction,we have designed an algorithm for recognizing suspicious financial transactions based on frequent item digging in data stream.The algorithm has improved the derogatory counting process and solved the excessive deletion problem in the data processing proce...