吉林大学学报（理学版）2024，Vol.62Issue(1)：116-121,6.DOI:10.13413/j.cnki.jdxblxb.2023181
基于多项式特征生成的卷积神经网络
Convolutional Neural Networks Based on Polynomial Feature Generation
摘要
Abstract
Based on the polynomial feature generation method for one-dimensional feature data,we proposed a data augmentation algorithm that used the polynomial feature generation method to generate feature data for high-dimensional feature data.At the same time,we proposed an algorithm that combined the generated polynomial feature data with the neural network model during convolutional neural network training,which could organically combine the generated polynomial feature data with the convolutional neural network model,and improve the low recognition accuracy and the limited generalization performance of model caused by data limitations such as limited data samples,fixed total number of data samples,and differences in available data samples when modeling convolutional neural network models.Experimental results show that the accuracy of the convolutional neural network model using this method achieves significant improvement.关键词
卷积神经网络/特征生成/多项式/特征堆叠Key words
convolutional neural network/feature generation/polynomial/feature stacking分类
信息技术与安全科学引用本文复制引用
刘铭,肖志成,于晓东..基于多项式特征生成的卷积神经网络[J].吉林大学学报（理学版）,2024,62(1):116-121,6.基金项目
吉林省发改委基本建设项目(批准号:2022C043-2)和吉林省自然科学基金(批准号:20200201157JC). （批准号:2022C043-2）
