Closed wewill7 closed 3 years ago
In the example on Neupy page they do normalization from the preprocessing in Skitlearn library:
import numpy as np
from sklearn import datasets, preprocessing
from sklearn.model_selection import train_test_split
from neupy import algorithms
**dataset = datasets.load_diabetes()
x_train, x_test, y_train, y_test = train_test_split(
preprocessing.minmax_scale(dataset.data),
preprocessing.minmax_scale(dataset.target.reshape(-1, 1)),
test_size=0.3,
)**
nw = algorithms.GRNN(std=0.1, verbose=False)
nw.train(x_train, y_train)
y_predicted = nw.predict(x_test)
mse = np.mean((y_predicted - y_test) ** 2)
mse
I want to know if the library has implemented the normalize for my input data? because when I change the 'std', I find it is very little is also ok despite my data is large. It troubles me . But I did not find the code that implements the normalization.