摘要
针对传统中医诊断信息技术中,存在样本不平衡及面色识别精度低的问题,提出了基于改进粒子群算法的支持向量机(SVM)模型。由于数据量的限制,选择SVM小样本分类模型。采用粒子群算法为SVM模型选择合理参数。采用惯性权重先增后减的更新策略,同时利用自适应高斯模型对粒子群算法进行改进,应对容易陷入局部最优的问题。研究了384个面部图像,并利用测试数据在训练模型上进行预测。实验结果显示,模型的平均识别精确率达92.26%,相比于传统的SVM算法高出5.25%。
To improve the process of traditional Chinese medical diagnosis with information technology, the imbalance of samples and the low accuracy of facial color recognition, an SVM model based on an improved particle swarm optimization(PSO) algorithm was proposed. Due to the limitation of the number of data, the SVM model was selected for small sample classification. PSO was used to choose appropriate parameters for the SVM model. In order to deal with the problem of easily falling into local optimum, the updating strategy of increasing inertia weight first and then decreasing it is adopted, and the PSO algorithm is improved by using the adaptive Gauss model. 384 facial images were studied, and the trained model was used to predict the testing data. Experimental results illustrate that the average recognition accuracy can reach 92. 26%, which is 5. 25% higher than the traditional SVM algorithm.
作者
李周姿
冯跃
林卓胜
徐红
LI Zhou-zi;FENG Yue;LIN Zhuo-sheng;XU Hong(Faculty of Intelligent Manufacturing,Wuyi University,Jiangmen Guangdong 529020,China;Victoria University,Melbourne Australia)
出处
《计算机仿真》
北大核心
2022年第4期241-247,共7页
Computer Simulation
基金
广东省五邑大学校内科研项目(2018TP023)
广东省五邑大学校内科研项目(2018GR003)。