OpenCV python slearnランダム超パラメータ検索の実現

"""
          sklearn       
"""
import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
import sklearn
import pandas as pd
import os
import sys
import tensorflow as tf
from tensorflow_core.python.keras.api._v2 import keras #      python
from sklearn.preprocessing import StandardScaler
from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import train_test_split, RandomizedSearchCV
from scipy.stats import reciprocal

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
assert tf.__version__.startswith('2.')

# 0.         
print(tf.__version__)
print(sys.version_info)
for module in mpl, np, sklearn, pd, tf, keras:
  print("%s version:%s" % (module.__name__, module.__version__))


#       
def plot_learning_curves(his):
  pd.DataFrame(his.history).plot(figsize=(8, 5))
  plt.grid(True)
  plt.gca().set_ylim(0, 1)
  plt.show()


# 1.      california   
housing = fetch_california_housing()

print(housing.DESCR)
print(housing.data.shape)
print(housing.target.shape)

# 2.                 
x_train_all, x_test, y_train_all, y_test = train_test_split(
  housing.data, housing.target, random_state=7)
x_train, x_valid, y_train, y_valid = train_test_split(
  x_train_all, y_train_all, random_state=11)

print(x_train.shape, y_train.shape)
print(x_valid.shape, y_valid.shape)
print(x_test.shape, y_test.shape)

# 3.      
scaler = StandardScaler()
x_train_scaled = scaler.fit_transform(x_train)
x_valid_scaled = scaler.fit_transform(x_valid)
x_test_scaled = scaler.fit_transform(x_test)


#   keras  
def build_model(hidden_layers=1, #       
        layer_size=30,
        learning_rate=3e-3):
  #      
  model = keras.models.Sequential()
  model.add(keras.layers.Dense(layer_size, activation="relu",
                 input_shape=x_train.shape[1:]))
 #      
  for _ in range(hidden_layers - 1):
    model.add(keras.layers.Dense(layer_size,
                   activation="relu"))
  model.add(keras.layers.Dense(1))

  #       
  optimizer = keras.optimizers.SGD(lr=learning_rate)
  model.compile(loss="mse", optimizer=optimizer)

  return model


def main():
  # RandomizedSearchCV

  # 1.   sklearn model
  sk_learn_model = keras.wrappers.scikit_learn.KerasRegressor(build_model)

  callbacks = [keras.callbacks.EarlyStopping(patience=5, min_delta=1e-2)]

  history = sk_learn_model.fit(x_train_scaled, y_train, epochs=100,
                 validation_data=(x_valid_scaled, y_valid),
                 callbacks=callbacks)
  # 2.       
  # f(x) = 1/(x*log(b/a)) a <= x <= b
  param_distribution = {
    "hidden_layers": [1, 2, 3, 4],
    "layer_size": np.arange(1, 100),
    "learning_rate": reciprocal(1e-4, 1e-2),
  }

  # 3.       
  # cross_validation:     n , n-1  ,       .
  random_search_cv = RandomizedSearchCV(sk_learn_model, param_distribution,
                     n_iter=10,
                     cv=3,
                     n_jobs=1)
  random_search_cv.fit(x_train_scaled, y_train, epochs=100,
             validation_data=(x_valid_scaled, y_valid),
             callbacks=callbacks)
  # 4.     
  print(random_search_cv.best_params_)
  print(random_search_cv.best_score_)
  print(random_search_cv.best_estimator_)

  model = random_search_cv.best_estimator_.model
  print(model.evaluate(x_test_scaled, y_test))

  # 5.        
  plot_learning_curves(history)


if __name__ == '__main__':
  main()