BP neural network derivation + Example

Basic concepts are not popularized .
Today, let's push it BP The basis of .
The original model diagram is as follows ：

A very simple model , Input and output ,O Represents the predicted value , purple t Represents the actual value .
This is the time , We see that the predicted results differ greatly from the actual values , Want to pass BP The neural network adjusts the parameters .

Pair coefficient sum bias The process of deriving . No loss of generality ,
below , For when j=0,k=0 This situation , Conduct case derivation .

The final result is shown in the figure below .