时序基因表达缺失值的加权双向回归估计算法

摘要

Due to the limited experimental condition, there are missing values in gene expression data which make the following use difficult. Estimating missing values without data destroy and information lost has become an important work of bio-information. By weighted kernel function, it can find out rows and columns having largest similar coefficient with the rows and columns containing missing values. An estimation method based on double weighted regression is introduced by using weighted kernel function. It makes the information data more abundant by considering gene space correlation and time correlation in regression. Comparing with other methods, the weighted double regression method can obtain better estimation result.%由于受实验条件等客观因素制约,实验所得到的基因表达谱数据存在数据缺失的现象,不利于数据的后续使用.如何在不丢失数据信息、不影响数据整体使用的情况下,对实验数据进行估计、填充已成为目前生物信息学研究的热点.本文通过利用核加权函数提取与缺失值所在的行列具有最大相似性的行列信息,提出了基于双向核加权回归估计的算法.在回归过程中同时考虑基因表达的空间相关性和时问相关性信息,使回归算法使用的信息更加充分.通过与其他缺失值估计算法相比较,加权双向回归算法的估计结果较好.