论文部分内容阅读
构建一种基于粒子群算法-支持向量机(PSO-SVM)的磁共振功能成像(fMRI)时间序列分类诊断模型,通过针对脑区多维时间序列数据的深层次分析实现病症患者和健康者的准确判断与区分,为面向fMRI时间序列数据的病症诊断和预测提供有效科学依据.该方法在以下4个方面不同于其他已有相关研究工作:(1)构建基于自回归模型的脑区多维时间序列数据特征表示;(2)构建基于支持向量机模型的脑区多维时间序列数据分类机制;(3)构建基于粒子群算法的分类学习参数寻优策略;(4)建立融合上述特征表示、优化分类与参数优选模式的fMRI时间序列数据分类诊断模型.通过以精神抑郁症作为实证分析的具体案例,所提出分类诊断模型已取得良好实验效果,展示出其有效性与合理性.
A time division diagnosis model of magnetic resonance imaging (fMRI) based on Particle Swarm Optimization-Support Vector Machine (PSO-SVM) was constructed to realize the accurate diagnosis of patients and healthy people by the deep analysis of multidimensional time series data And provides an effective scientific basis for the diagnosis and prediction of the illness for the fMRI time series data.This method is different from other existing related research work in the following four aspects: (1) Constructing multi-dimensional time series based on autoregressive model Data feature representation; (2) constructing a brain multidimensional time series data classification mechanism based on support vector machine model; (3) building a classification strategy based on particle swarm optimization algorithm optimization strategy; (4) establishing a fusion of the above feature representation, And parameter optimization mode of fMRI time series data classification diagnostic model.Diagnosis of depression as a specific case of empirical analysis, the proposed classification diagnostic model has achieved good experimental results show its effectiveness and rationality.