Hey Sunita,
When testing an ANN with new input data, it is essential to normalize the input data to ensure that all input features are on a similar scale. This normalization process will help the ANN to learn more effectively and improve its performance. You can use MATLAB's mapminmax function to normalize the input data to a specific range, such as [-1, 1] or [0, 1].
For example, suppose you have a matrix X containing your input data. You can normalize it using the following code:
X_normalized = mapminmax(X, 0, 1);
This code will normalize the input data in X to the range [0, 1]. You can adjust the range by changing the second and third arguments of the mapminmax function.
There are other normalization techniques besides mapminmax that you can use in MATLAB. Here are a few examples:
- zscore: This function normalizes the data to have zero mean and unit variance. It is useful when the data has a Gaussian distribution.
- rescale: This function scales the data to a specified range. It is similar to mapminmax, but it allows you to specify the output range directly.
- normalize: This function normalizes the data to have unit norm. It is useful when you want to compare the magnitude of different feature vectors.
You can refer to the official Mathworks Documentation for thorough understanding
I hope this helps!
