一种基于朴素贝叶斯的银行卡批量业务预测方法

摘要

目前银行卡批量业务已成为一种比较常见且重要的金融产品,其交易来源于收单机构以结算文件的形式批量上送,对后台处理系统的性能要求较高.基于朴素贝叶斯理论提出了银行卡批量业务日最大TPS(每秒交易笔数)的预测方法.利用生产数据在测试环境的回放,得到测试环境最大TPS与生产实际值的回归关系,进而在测试环境拟合得到大量样本数据作为训练集,并取其中的20％数据作为测试集,实际验证了该方法具有准确度高、鲁棒性强的特点.%The bank card batch service has become a more commom and important financial product.The transaction comes from the acquire institution in the form of batch delivery.So high performance requirements for background processing system are proposed.This paper proposes a bank card batch service max TPS (tranction per second) prediction method.We use the production data to be played back in the test environment.And the training set is collected by the test result,twenty percent of the data is used to be the training set.Native Bayes method is introduced to train the classifier.Experiments show that our method has the characteristics of high accuracy and strong robustness.