摘要:近年来,中国债券市场稳健发展。为避免异常交易而产生的损失,用户通过债券估值来寻找合理的债券价格。债券估值主要通过收益率曲线变化体现,收益率曲线则根据大量的市场数据和模型公式构造。要让用户及时得知债券变化就必须提高计算速度。SparkR是将Spark和R结合,Spark的分布式并行计算则可以快速提高计算速度,R语言的建模可以筛选有效的数据。基于SparkR的债券估值系统不仅可以帮助用户维护管理债券信息,还可以通过曲线图让用户实时直观地了解债券行情和未来趋势,其自定义个性化的估值可以满足用户不同需求。论文介绍了基于SparkR的债券估值系统的功能分析与实现方法,重点介绍债券估值的管理维护以及基于SparkR的计算,并为此类系统提供了技术参考,最后对系统的改进和发展提出了建议和展望。43765
Design and implementation of bond valuation system based on SparkR
Abstract: In recent years, China's bond market is developing steadily. In order to avoid losses arising from unusual transactions, customers should find a reasonable bond prices by using bond valuations. The bond valuations reflect changes in the yield curve, and the curve is based on a lot of market data and the model formula structure. The method to alert customers change the bonds in time is improving computing speed.SparkR is a combination of Spark and R, Spark distributed parallel computing can quickly speed up calculations, modeling R language can filter valid data. Based on SparkR bond valuation system can not only help you maintain debt management information, but also it can graph lets users intuitively understand the real-time bond market and future trends, its custom personalized estimate to meet your needs.This thesis introduced based on SparkR bond valuation system function analysis and realization method, focuses on the bond valuation management maintenance and based on the calculation of SparkR and for such a system provides technical reference. Finally, the improvement and development of the system put forward suggestions and prospects.
Keywords: Bond Valuation;SparkR;Management System
目录
摘要 i
Abstract i
目录 ii
1 引言 1
1.1 债券估值发展现状 1
1.1.1 估值曲线分类 1
1.1.2 国外债券估值的发展情况 1
1.1.3 国内债券估值的发展情况 2
1.2 债券实时估值存在的问题 3
1.3 本文的主要内容 4
1.4 本文的篇章结构 4
2 债券实时估值核心技术 5
2.1 样条函数的理论基础 5
2.2 基于R语言模型集成 6
2.3 基于SPARK的分布式运算以及数据共享 7
3 债券实时估值系统需求分析 10
3.1 外部源管理模块 10
3.1.1 交易所数据管理 11
3.1.2 外汇交易中心数据管理 11
3.2 债券估值信息管理 11
3.2.1 债券估值信息维护