중간층 만들기(값예측)

data : iris
X : iris.data[:,[0,1,2]]
y : iris.data[:,3]
w : random.randn
중간층의 활성함수 : Relu

- 중간층 뉴런 10개를 포함해 y 값 추정하기.

 

 

import numpy as np
from sklearn.datasets import load_iris

X=iris.data[:,[0,1,2]]
y=iris.data[:,3]

w=np.random.randn(3,10)
b=np.random.randn(10)

# Relu 적용
u=np.maximum(0,X@w+b)

ww=np.random.randn(10,1)
b=np.random.randn(1)

pred_y=u@ww+b
pred_y

 

array([[-13.68528237],
       [-11.99422824],
       [-12.98971925],
       [-12.7758113 ],
       [-14.1480654 ],
       [-14.70377861],
       [-13.92116272],
       [-13.33053468],
       [-12.318053  ],
       [-12.29810617],
       [-14.09602274],
       [-13.6442963 ],
       [-12.08320838],
       [-12.73788397],
       [-15.07128567],
       [-16.44570249],
       [-14.98347842],
       [-13.68528237],
       [-14.06303528],
       [-14.73676606],
       [-12.83476419],
       [-14.36296318],
       [-14.78368579],
       [-12.72790174],
       [-13.90478229],
       [-11.76539819],
       [-13.26060972],
       [-13.52637727],
       [-13.22249934],
       [-13.04810479],
       [-12.41211623],
       [-12.97461409],
       [-15.76919457],
       [-15.94598197],
       [-12.29810617],
       [-12.79270377],
       [-13.39928674],
       [-12.29810617],
       [-12.64195103],
       [-13.24155453],
       [-13.84418746],
       [ -9.80345359],
       [-13.46340441],
       [-13.63441261],
       [-14.87190345],
       [-12.08320838],
       [-14.66684111],
       [-13.09970934],
       [-14.18500289],
       [-13.02665674],
       [-10.55877029],
       [-11.73668418],
       [-10.54696279],
       [ -9.63335747],
       [ -9.95534421],
       [-11.6706105 ],
       [-12.54633007],
       [-10.78785475],
       [-10.14080903],
       [-11.86522118],
       [ -9.09334343],
       [-11.78105416],
       [ -8.09632143],
       [-11.35394703],
       [-11.52514111],
       [-10.56334322],
       [-12.71732575],
       [-10.68730729],
       [ -8.079941  ],
       [-10.14272032],
       [-13.12347951],
       [-10.33541972],
       [ -9.43417385],
       [-10.94322034],
       [-10.33084678],
       [-10.3778784 ],
       [ -9.45321592],
       [-10.6735885 ],
       [-11.40555158],
       [ -9.9808705 ],
       [ -9.87042684],
       [ -9.78359818],
       [-10.51364997],
       [-11.10507016],
       [-13.16784949],
       [-13.45918504],
       [-10.82382921],
       [ -8.17857716],
       [-12.37001111],
       [-10.45481085],
       [-11.21285219],
       [-11.67784507],
       [-10.18975194],
       [-10.15186619],
       [-11.22465969],
       [-12.2315779 ],
       [-11.82085121],
       [-10.78137052],
       [-10.48757172],
       [-11.32329585],
       [-13.67510266],
       [-11.5555939 ],
       [-10.55399897],
       [-11.68488124],
       [-11.81874152],
       [-10.03549025],
       [-12.24052537],
       [-10.04006318],
       [ -9.31458432],
       [-12.96675458],
       [-12.03239428],
       [-10.37768001],
       [-10.88246993],
       [-10.87257372],
       [-11.96632059],
       [-12.43131347],
       [-11.55825554],
       [-13.18287058],
       [ -8.42780759],
       [ -8.96460804],
       [-11.65231877],
       [-12.24318701],
       [ -9.07560366],
       [-10.25562723],
       [-12.5135692 ],
       [-11.23701914],
       [-10.80478713],
       [-11.93833105],
       [-11.04889268],
       [-10.24190844],
       [ -9.2304173 ],
       [-12.47186086],
       [-11.04889268],
       [-10.84001125],
       [-10.90322491],
       [ -9.37608508],
       [-13.73851471],
       [-12.1942441 ],
       [-12.07676426],
       [-10.98110609],
       [-11.60528715],
       [-10.72062011],
       [-11.5555939 ],
       [-12.05123796],
       [-12.5135692 ],
       [-10.84724582],
       [ -9.52100251],
       [-11.29776956],
       [-13.79011925],
       [-12.56251211]])

 

mse=((y-pred_y)**2).mean()
mse

173.90711513080973