College of Computer Science and Artificial Intelligence, Wenzhou University, Wenzhou, 325035, China.
Wenzhou Medical University, Wenzhou, China.
Comput Biol Med. 2022 Jun;145:105435. doi: 10.1016/j.compbiomed.2022.105435. Epub 2022 Apr 2.
Systemic lupus erythematosus is a chronic autoimmune disease that affects the kidney in most patients. Lupus nephritis (LN) is divided into six categories by the International Society of Nephrology/Renal Pathology Society (ISN/RPS). The purpose of this research is to build a framework for discriminating between ISN/RPS pure class V(MLN) and classes III ± V or IV ± V (PLN) using real clinical data. The framework is developed by merging a hybrid stochastic optimizer, moth-flame algorithm (HMFO), with a support vector machine (SVM), dubbed HMFO-SVM. The HMFO is constructed by enhancing the original moth-flame algorithm (MFO) with a bee-foraging learning operator, which guarantees that the algorithm speeds convergence and departs from the local optimum. The HMFO is used to optimize parameters and select features simultaneously for SVM on clinical SLE data. On 23 benchmark tests, the suggested HMFO method is validated. Finally, clinical data from LN patients are analyzed to determine the efficacy of HMFO-SVM over other SVM rivals. The statistical findings indicate that all measures have predictive capabilities and that the suggested HMFO-SVM is more stable for analyzing systemic LN. HMFO-SVM may be used to analyze LN as a feasible computer-assisted technique.
系统性红斑狼疮是一种慢性自身免疫性疾病，大多数患者的肾脏都会受到影响。狼疮肾炎（LN）根据国际肾脏病学会/肾脏病理学会（ISN/RPS）的分类标准可分为六类。本研究旨在利用真实临床数据建立一种框架，用于区分 ISN/RPS 纯五类（MLN）和三类加五类或四类加五类（PLN）。该框架是通过将混合随机优化器—— moth-flame 算法（HMFO）与支持向量机（SVM）合并而构建的，称为 HMFO-SVM。HMFO 通过增强原始 moth-flame 算法（MFO）的蜜蜂觅食学习算子来构建，这可以确保算法加快收敛速度并避免陷入局部最优。HMFO 用于同时优化 SVM 的参数和选择特征，从而对临床 SLE 数据进行优化。在 23 个基准测试中，验证了所提出的 HMFO 方法。最后，分析 LN 患者的临床数据，以确定 HMFO-SVM 相对于其他 SVM 竞争对手的疗效。统计结果表明，所有指标均具有预测能力，并且所提出的 HMFO-SVM 更适合分析系统性 LN。HMFO-SVM 可作为一种可行的计算机辅助技术用于分析 LN。
