How to preprocess the data before ANN model training

Acutally, I'm doing the ANN modeling. Because the Ids has large dynamic range from 4.1E-16 to 6.35E-6. So, I've to preproecess the Ids data for accurancy improvment. And I plan used the preproecess that paper mentioned. But, I don't know how to conduct it (Transfer the measurement Ids to Ids').

Ids = [6.35E-06
6.18E-06
6.01E-06
5.84E-06
5.68E-06
5.52E-06
5.36E-06
5.20E-06
5.05E-06
4.89E-06
4.75E-06
4.60E-06
4.45E-06
4.31E-06
4.17E-06
4.03E-06
3.90E-06
3.76E-06
3.63E-06
3.50E-06
3.38E-06
3.25E-06
3.13E-06
3.01E-06
2.90E-06
2.78E-06
2.67E-06
2.56E-06
2.45E-06
2.35E-06
2.25E-06
2.15E-06
2.05E-06
1.95E-06
1.86E-06
1.77E-06
1.68E-06
1.60E-06
1.51E-06
1.43E-06
1.35E-06
1.28E-06
1.20E-06
1.13E-06
1.06E-06
9.95E-07
9.30E-07
8.68E-07
8.08E-07
7.50E-07
6.95E-07
6.41E-07
5.90E-07
5.41E-07
4.95E-07
4.51E-07
4.09E-07
3.69E-07
3.31E-07
2.95E-07
2.62E-07
2.31E-07
2.02E-07
1.75E-07
1.50E-07
1.27E-07
1.06E-07
8.77E-08
7.09E-08
5.60E-08
4.31E-08
3.20E-08
2.28E-08
1.52E-08
9.39E-09
5.11E-09
2.28E-09
7.14E-10
1.12E-10
6.28E-12
2.10E-13
6.54E-15
3.27E-16
-2.03E-16
-2.06E-16
4.78E-16
3.29E-16
-3.03E-16
4.10E-16];

Answer 1

[Udit06]

Hi Chuanjung,

I understand that you need help to preprocess to using the preprocessing function proposed in the paper. You can follow the following steps to achieve the same:

1) Define a function say “my_preprocessing”, that takes as input and returns as an output using the equation given in the research paper.

2) Use “finverse” function given in MATLAB to find the inverse of the function “my_processing”. You can refer to below MathWorks documentation to understand more about the “finverse” function.

https://www.mathworks.com/help/symbolic/sym.finverse.html

3) Use the inverse function obtained in step 2 to preprocess to .

I hope this helps.