基于模糊神经网络的金融仓单风控数据筛选系统  

Financial Warehouse Receipt Risk Control Data Screening System Based on Fuzzy Neural Network

作　　者：顾凌云 GU Ling-yun(Shanghai Icekredit,Inc.,Shanghai 200120 China)

机构地区：[1]上海冰鉴信息科技有限公司,上海200120

出　　处：《自动化技术与应用》2025年第3期74-78,共5页Techniques of Automation and Applications

基　　金：上海市软件和集成电路产业发展专项资金项目(200316)。

摘　　要：现有的数据筛选系统使用方法实时性较低,影响结果置信度,无法满足特定场景下金融仓单风控的需要。因此设计一种基于模糊神经网络的风控数据筛选系统。将仓单页面表示为二维矩阵,经过离散余弦变换将仓储量和仓单类型进行转换,实现金融仓单页面精准分析,对金融仓单中数据进行特征提取,利用数据集中分类的属性方程对风控数据进行特征向量匹配,分析模糊神经网络的输入和输出数据,对所有的数据进行遍历,得到风控数据结果。在系统性能测试中,设计的系统输出的数据具有很强的特征聚类性。The existing data filtering system has low real-time performance,which affects the confidence of the results and cannot meet the needs of risk control of financial warehouse receipts in specific scenarios.Therefore,this paper designs a risk control data screening system based on fuzzy neural network.The warehouse receipt page is represented as a two-dimensional matrix,and the warehouse reserves and warehouse receipt types are converted through discrete cosine transformation to achieve accurate analysis of the financial warehouse receipt page,feature extraction of the data in the financial warehouse receipt,feature vector matching of the risk control data using the attribute equation classified in the data set,analysis of the input and output data of the fuzzy neural network,and traversal of all data to obtain the risk control data results.In the system performance test,the output data of the designed system has strong feature clustering.

关 键 词：模糊神经网络 金融仓单 数据筛选 特征匹配 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.3[自动化与计算机技术—控制科学与工程]