BackPropagation_02_NeuralNetworkAlgorithmDerivation_and_CodeImplementationNotes

The Unknown Word

The First Word	The Second Word
derivation	[dere'veition]求导
fraction	分数
cross-entropy	交叉熵
entropy	['entrepi]熵
feedforward	['fid'fo:werd]前馈
quadratic	二次的[kwa:'draetik]

Neuron

What is a neuron？ from the point of view of function image, this is a straight line.

What is the use of a neuron? To illustrate the use, we must give an application scenario: classification. A neuron is a straight line, equivalent to the Chu River Han boundary, can be divided between red chess and green chess, at this time it is a classifier. Therefore, in linear scenarios, a single neuron can achieve the function of classification, it can always learn a suitable line to distinguish the two types of elements.

Look at the effect picture,Transmission

对上面的图简单说明一下:(x1,x2) 对于神经元的输入都是 x, 而对我们而言，这数据就是意义上的点的坐标，我们习惯写成 (x,y)。

TO BE CONTINUED

Reference

Transmission_MachineLearningIntroduction

转载于:https://www.cnblogs.com/hugeng007/p/9535811.html