基于人工智能的结直肠癌分类识别研究

基于人工智能的结直肠癌分类识别研究

摘要：随着我国人口老龄化和人工智能系统的发展，目前行业内相对比较传统的结直肠癌分类方法无法满足客户的需求。为了改进传统结直肠癌分类方法在结直肠癌分类上得短板问题，同时结直肠癌的分类识别方法目前已有先进且稳定的水平，但由于结直肠癌的原理和组成相对复杂，分类难度相对比较高，因此，在本次毕业设计中，将使用相对先进、快捷、智能的分类机制。本次设计采用线性判别分析（LDA）的方法进行结直肠癌的数据降维和分类识别。

关键词：人工智能；结直肠癌；分类识别；线性判别分析；SVM

Research on classification and recognition of colorectal cancer based on artificial intelligence

Abstract: With the aging of China's population and the development of artificial intelligence systems, the current relatively traditional colorectal cancer classification methods in the industry cannot meet the needs of customers. In order to improve traditional colorectal cancer classification method were short in the classification of colorectal cancer, colorectal cancer classification recognition method at the same time there are advanced and stable level, but due to the principle and composition of colorectal cancer is relatively complex, classified difficulty is relatively high, therefore, in this graduation design, will use relatively advanced, efficient, intelligent classification system.  In this design, linear discriminant analysis (LDA) was used for data reduction and classification recognition of colorectal cancer.

Keywords: Artificial intelligence； Colorectal cancer；Classification recognition；LDA；SVM

目录

目录..................................................................................................... ii

1 绪论................................................................................................. 1

1.1 研究目的与意义....................................................................... 1

1.2 结直肠癌分类识别技术的现状与水平................................... 1

1.3 结直肠癌分类识别技术的发展趋势....................................... 3

1.4 课题研究内容及章节安排....................................................... 3

2 相关理论与技术............................................................................. 4

2.1 支持向量机SVM..................................................................... 4

2.1.1 算法原理.......................................................................... 4

2.1.2 核心思想.......................................................................... 4

2.2 线性判别分析LDA的定义与原理........................................ 4

2.2.1 线性判别分析的定义...................................................... 4

2.2.2 瑞丽熵原理...................................................................... 5

2.2.3 LDA二类降维原理.......................................................... 6

2.2.4 LDA多类别降维原理...................................................... 7

2.3 LDA的数据降维的可视化...................................................... 8

2.4 数据集的划分........................................................................... 9

3 结直肠癌分类识别预测模型....................................................... 10

3.1 病理分析................................................................................. 10

3.1.1 磁共振图像.................................................................... 10

3.1.2 医学依据........................................................................ 10

3.2 需求分析................................................................................. 11

3.3 模型结构................................................................................. 11

3.4 诊断分类................................................................................. 12

3.4.1 分类................................................................................ 12

3.4.2 集成分类........................................................................ 12

4 实验结果分析............................................................................... 13

4.1 LDA的实验结果.................................................................... 13

4.1.1 降低到一维的分类效果................................................ 13

4.1.2 降低到二维的分类效果................................................ 14

4.1.3 降低到三维分类效果.................................................... 15

4.2 多元逻辑回归的实验结果..................................................... 16

4.2.1 降低到一维的分类效果................................................ 16

4.2.2 降低到二维的分类效果................................................ 18

4.2.3 降低到三维分类效果.................................................... 19

4.3 SVM的实验结果.................................................................... 21

4.3.1 降低到一维的分类效果................................................ 21

4.3.2 降低到二维的分类效果................................................ 21

4.3.3 降低到三维分类效果.................................................... 22

5 结束语........................................................................................... 24

致谢................................................................................................... 25

参考文献........................................................................................... 26