基于云模型癌症相关基因分类预测的研究

发布时间：2018-04-05 20:19

本文选题：云模型理论　切入点：粒子群优化算法　出处：《吉林大学》2012年硕士论文

【摘要】：随着科学技术的不断发展与后基因组时代的来临，人类对基因的了解越发深入，同时也大大提高了基因表达数据的检测手段与检测技术，使研究人员能在较短的时间与较少的实验次数下获得大量的基因表达数据。这些数据对于研究各种疾病的发病机理、疾病的诊断、以及开发新型药物和对疾病在基因水平进行基因治疗都具有重要意义。癌症是影响人类健康的主要疾病。在基因水平上对癌症相关基因进行分类和预测研究是了解癌症的发病机理，找到基因表达数据的变化与癌症病理特征之间的关系，从而开发出针对特定基因的新型药物对癌症进行治疗的关键步骤。然而在海量的基因数据中只有较少的样本可以进行研究分析，这就造成了严重的“维灾”现象，同时因为癌症基因数据中的大量沉余，导致了分类性能和准确性的严重下降。为了解决上述问题，本文将使用基于云模型的分类器对癌症相关基因进行分类研究，，目前应用云模型理论对癌症相关基因进行分类的相关文献尚不多见，本文意在利用云模型理论在数据挖掘方面的优势，结合粒子群优化算法，对癌症相关基因进行分类预测研究。本文主要工作如下：（1）详细对生物信息学（bioinformatics）进行总结与阐述，包括生物信息学（bioinformatics）的定义、产生与发展、研究领域和近期研究的主要成就等。（2）对当前生物信息学研究的热点问题——癌症相关基因的分类与预测问题进行分析与研究。主要包括癌症的发生和发展与细胞周期之间的规律、本文所应用的数据集中的特征基因及其生物学意义以及国内外癌症相关基因的研究进展情况。（3）对云模型理论进行详细的阐述，包括云的定义、云模型的基本特点和云模型的三个基本数字特征。分析与讨论了云模型的发生器及其相关算法，并对近几十年来云模型理论的研究进展情况进行介绍。（4）将粒子群优化算法与云模型理论相结合，应用云模型分类器对癌症相关基因进行分类预测研究，将基于云模型的分类器与其他有类似功能的分类方法进行比较研究，分析各自的优缺点，并提出改进方案。同时分析研究各种应用不同算法的云分类器在分类效果与分类效率上的不同，对其进行比较，验证了基于粒子群云模型癌症相关基因分类预测的有效性。
[Abstract]:With the continuous development of science and technology and the advent of post-genome era, the understanding of genes has become more and more in-depth, and the detection methods and techniques of gene expression data have also been greatly improved.This allows researchers to obtain large amounts of gene expression data in a shorter time and fewer experiments.These data are of great significance for the study of the pathogenesis and diagnosis of various diseases, as well as the development of new drugs and gene therapy for diseases at the gene level.Cancer is a major disease affecting human health.Classification and prediction of cancer-related genes at the gene level is to understand the pathogenesis of cancer and to find out the relationship between the changes of gene expression data and the pathological characteristics of cancer.A key step in cancer treatment is to develop new drugs for specific genes.However, only a small number of samples can be studied and analyzed in a large amount of genetic data, which results in a serious "disaster of maintenance" phenomenon. At the same time, because of the large amount of residual in cancer gene data, the classification performance and accuracy are seriously reduced.In order to solve the above problems, this paper will use cloud model-based classifier to classify cancer related genes. At present, there are few literatures about cancer related genes classification based on cloud model theory.This paper aims to make use of the advantage of cloud model theory in data mining, combining with particle swarm optimization algorithm, to study the classification and prediction of cancer related genes.The main work of this paper is as follows:1) summarize and expound bioinformatics in detail, including the definition, production and development of bioinformatics, the research field and the main achievements of recent research, etc.This paper analyzes and studies the classification and prediction of cancer related genes, which is a hot topic in bioinformatics.It mainly includes the regularity between the occurrence and development of cancer and cell cycle, the characteristic genes and their biological significance in the data set used in this paper, and the research progress of cancer related genes at home and abroad.3) the theory of cloud model is expounded in detail, including the definition of cloud, the basic characteristics of cloud model and the three basic numerical features of cloud model.The generator of cloud model and its related algorithms are analyzed and discussed, and the research progress of cloud model theory in recent decades is introduced.4) combining particle swarm optimization algorithm with cloud model theory, applying cloud model classifier to classify and predict cancer related genes, comparing the classifier based on cloud model with other classification methods with similar functions.The advantages and disadvantages of each are analyzed, and the improvement scheme is put forward.At the same time, the different classification effects and classification efficiency of different cloud classifiers with different algorithms are analyzed and compared to verify the effectiveness of the classification prediction of cancer related genes based on particle swarm cloud model.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP3;R730.2

【参考文献】