摘要
该文从几何的视角来看待感知机和神经网络,采用两个简单、便于演示的例子来解释感知机和BP神经网络的分类功能,并结合Python动画演示其权值、偏置值在训练中逐步收敛的过程。本文介绍的方法在教学实践中收到了较好的效果,可以帮助学生对感知机和BP神经网络产生感性认识。
In this paper, we perceive the perceptron and BP neural network from a geometric perspective. We explain the classifica-tion function of perceptron and neural network with two examples that simple and easy to visualize. We also use python animationto demonstrate how the weights and biases converge gradually. The method introduced in this paper has good effect in the teachingpractice, and can help students to build perceptual knowledge of the perceptron and the BP neural network.
作者
曹建立
赖宏慧
徐世杰
CAO Jian-li1, LAI Hong-hui2, XU Shi-jie3 (1. College of Mathematics and Science, Luoyang Normal University, Luoyang 471934, China; 2. School of Information Engineering, Gannan Medical University, Ganzhou 341000, China; 3. Luoyang Experimental Primary School, Luoyang 471000, China)
出处
《电脑知识与技术》
2018年第7期178-180,共3页
Computer Knowledge and Technology
基金
国家自然科学基金(批准号:61572246)
河南省科技创新人才支持计划(批准号:164100510003)