|
| 基于SCSO-SVM的结肠腺癌诊断研究 |
| Research on Colon Adenocarcinoma Diagnosis Based on SCSO-SVM |
| 投稿时间：2025-12-14 修订日期：2026-02-08 |
| DOI： |
| 中文关键词: 结肠腺癌 沙猫群算法 诊断模型 支持向量机 |
| 英文关键词: colon adenocarcinoma Sand Cat Swarm Optimization diagnostic model Support Vector Machine |
| 基金项目:国家自然科学基金项目（面上项目，重点项目，重大项目） |
|
| 摘要点击次数: 218 |
| 全文下载次数: 0 |
| 中文摘要: |
| 针对现有生物信息学对结肠腺癌早期诊断的技术尚不完善，难以在疾病早期诊断病情的问题，提出建立沙猫群算法的支持向量机结肠腺癌诊断模型（SCSO-SVM）的方法。该方法针对结肠腺癌患者肿瘤与非肿瘤组织中基因表达数据，采用差异表达分析与加权基因共表达网络获得共享基因，并对共享基因做功能富集分析。同时构建蛋白质相互作用网络，结合最大集团中心性和最大邻域分量两种算法得出核心基因。然后用最小绝对收缩和选择算子和随机森林得到特征基因，建立SCSO-SVM模型，并与灰狼算法、粒子群算法、遗传算法和传统支持向量机模型进行对比。SCSO-SVM在训练集、测试集和验证集上准确率分别为0.9942、0.9865、0.8235，AUC分别为0.99972、0.99918、0.86159，优于其它模型。说明该模型在结肠腺癌的预测具有更好的诊断效果。 |
| 英文摘要: |
| Due to the limitations of existing bioinformatics techniques in the early diagnosis of colon adenocarcinoma, which posed challenges in detecting the disease at its initial stages, a diagnostic model based on a Sand Cat Swarm Optimized Support Vector Machine (SCSO-SVM) was proposed. Initially, gene expression data from tumor tissues and adjacent non-tumor tissues of colorectal adenocarcinoma patients were analyzed. Differential expression analysis and weighted gene co-expression network analysis (WGCNA) were applied to screen for shared genes. Functional enrichment analysis was performed to explore the biological functions of these shared genes. Meanwhile, a protein-protein interaction (PPI) network was constructed based on shared genes, and core genes were identified by integrating maximum cluster centrality and maximum neighborhood component algorithms. Feature genes were further extracted by using the least absolute shrinkage and selection operator (LASSO) and random forest methods. Feature genes were then used to develop the SCSO-SVM diagnostic model. To evaluate its performance, the proposed SCSO-SVM model was compared with four other models: grey wolf optimization (GWO)-SVM, particle swarm optimization (PSO)-SVM, genetic algorithm (GA)-SVM, and the traditional SVM. Results showed that the accuracy rates of SCSO-SVM on the training set, test set and validation set were 0.9942, 0.9865 and 0.8235 respectively, and the AUC values were 0.99972, 0.99918 and 0.86159 respectively. These performance metrics outperformed those of the other comparative models. It indicated that the SCSO-SVM model possessed superior diagnostic efficacy for the early prediction of colorectal adenocarcinoma. |
|
View Fulltext
查看/发表评论 下载PDF阅读器 |
| 关闭 |
分享按钮