Neural Networks: Data normalization

■ Neural Networks: Data normalization

http://www.cs.sunysb.edu/~cse634/ch6NN.pdf

뉴럴 네트워크에서는 입력과 출력 데이터의 정규화가 필요하다. 위 사이트에 보면 데이터 정규화 방법으로 두개가 나와 있다. 

All values of attributes in the dataset has to be changed to contain values in the interval [0,1], or [-1,1].

모든 데이터는 0~1, -1~1 사이 값에 포함되도록 바꿔야 한다. 

Two basic normalization techniques:

- Max-Min normalization

- Decimal Scaling normalization 

Max-Min normalization formula

Example: We want to normalize data to range of the interval [0,1].

We put: new_max A=1, new_min A  = 0

Say, max A was 100 and min A was 20(That means maximum and minimum values for the attribute A).

Now, if v=40(if for this particular pattern, attribute value is 40), v' will be calculated as, 

v'=(40-20)*(1-0)/(100-20)+0

   = (20 * 1)/80

   = 0.4(?) 아닌거 같은데... 0.25 아닌가?!!

Decimal Scaling Normalization formula

where j is the smallest integer such that max|v’|<1.

Example :

A–values range from -986 to 917. Max |v| = 986.

v = -986 normalize to v’= -986/1000 = -0.986