摘要
交通事故救助基金数据具有可分离性低、易导致模型过度拟合等特征。为减少上述因素的影响,提出了基于深度学习模型的交通事故救助基金数据聚类方法。该方法首先采用模糊C均值算法预处理交通事故救助基金数据;其次构建基于卷积神经网络的深度学习模型,提高数据可分离性;最后利用非线性神经元映射函数求解模型,完成交通事故救助基金数据聚类。实验结果表明,该方法的数据重采样率较低、数据聚类效果优且优化精度较高。
The data of traffic accident rescue fund has the characteristics of low separability and is prone to causing model over-fitting.In order to reduce the influence of the above factors,a data clustering method of traffic accident relief fund based on deep learning model is proposed.Firstly,the fuzzy C-means algorithm is used to preprocess the data of traffic accident rescue fund,and then a deep learning model based on convolutional neural network is constructed to improve the separability of data.Finally,the nonlinear neuron mapping function is used to solve the model and complete the data clustering of traffic accident rescue fund.The experimental results show that the data resampling rate of this method is low,the data clustering effect is good and the optimization accuracy is high.
作者
朱璇
ZHU Xuan(Guizhou University People’s Armed Forces College,Guiyang Guizhou 550000,China)
出处
《信息与电脑》
2025年第18期50-52,共3页
Information & Computer
关键词
卷积神经网络
非线性神经元映射函数
数据聚类
convolutional neural network
nonlinear neuron mapping function
data clustering
