【毕业论文】贝叶斯网络在试管婴儿胚胎移植数目的研究.doc

贝叶斯网络在试管婴儿胚胎移植数目的研究 摘要 试管婴儿技术历时30多年发展，已卓有成效。部分生殖中心的新鲜胚胎移植成功率已经达到50%以上，但由于试管婴儿的多胚胎移植，使得试管婴儿治疗出现另一问题--多胎妊娠。多胎妊娠会影响新生儿发育，增加孕妇生育风险，应该避免。单胚胎移植可以拒绝多胎妊娠，可这又会影响患者怀孕成功几率，导致治疗失败。因此，如何在不降低妊娠率的前提下减少多胎发生一直备受瞩目。 为了发现试管婴儿技术中胚胎移植数目与多胎妊娠之间的规则，本文主要以数据挖掘理论为基础，运用贝叶斯信念网络的思想对试管婴儿技术进行研究。首先对数据挖掘技术做了概况性的阐述，介绍数据挖掘的方法、分类以及数据挖掘的应用和发展趋势；其次介绍了贝叶斯网及其相关技术：详细描述了贝叶斯网基础、参数估计、结构学习。最后，咨询领域专家建立基于试管婴儿治疗的网络结构模型，以MATLAB数据软件编写贝叶斯网参数学习的算法得到参数模型，发现试管婴儿技术中胚胎移植数目与孕妇最终怀孕后的胎数之间的关系。以VFP数据库管理系统编写代码，在现有的辅助生殖管理系统中增加以参数模型为基础的确定胚胎移植数目的功能。帮助医护人员选择在不降低治疗成功率前提下控制多胎发生的最合适的胚胎移植数目。 关键词：数据挖掘，贝叶斯网络，多胎妊娠，胚胎移植 Abstract Assisted reproductive technology which lasted more than 30 years development has been fruitful. The success rate of part of the center for fresh embryo transfer has been increased over 50%, however, this is accompanied by the emergence of another problem-multiple pregnancy. The excess transplantation embryos is the main cause of multiple pregnancies. How to reduce the multiple birth rate and maintain the pregnancy rate is always under attention. In order to find the rules between the number of embryo transfer in assisted reproductive technology and multiple pregnancy, this paper studies the Bayesian network used in assisted reproductive technology in data mining. First, it is described the data mining technology, introduced the methods of data mining and classified the application and trends of the data mining. Followed by a detailed description of the Bayesian network, based Bayesian network, parameter estimation, structure learning. Finally, consult an expert to establish the network structure, prepared by the use of MATLAB algorithm for mining the hidden information in the historical data, and add the information to the existing assisted reproductive management system to help medical staff to determine the number of embryos transferred. Key Words：Date mining, Bayesian networks, multiple pregnancy, embryo transfer目录 摘要 i Abstract ii 图目录 III 表目录 IV 第1章 绪论 1 1.1 引言 1 1.2 课题来源 1 1.3 试管婴儿技术简介 2 1.4 论文的研究内