实验三坤气学习

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import warnings

warnings.filterwarnings(action='ignore')
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
from sklearn.model_selection import train_test_split
import sklearn.neural_network as nn

data = pd.read_table('邮政编码数据.txt', sep=' ', header=None)
X = data.iloc[:, 1:-1]
Y = data.iloc[:, 0]

np.random.seed(1)
indices = np.random.choice(len(Y), 25)
plt.figure(figsize=(8, 8))
for i, idx in enumerate(indices):
img = np.array(X.iloc[idx, ]).reshape((16, 16))
plt.subplot(5, 5, i + 1)
plt.imshow(img, cmap=cm.gray_r)
plt.show()

X_train, X_test, Y_train, Y_test = train_test_split(X, Y, train_size=0.60, random_state=123)

hidden_nodes = np.arange(1, 20, 2)
activation_functions = ['relu', 'logistic']
train_err = np.zeros((len(hidden_nodes), 2))
test_err = np.zeros((len(hidden_nodes), 2))

for i, node in enumerate(hidden_nodes):
for j, act_func in enumerate(activation_functions):
model = nn.MLPClassifier(hidden_layer_sizes=(node,), activation=act_func, random_state=1)
model.fit(X_train, Y_train)
train_err[i, j] = 1 - model.score(X_train, Y_train)
test_err[i, j] = 1 - model.score(X_test, Y_test)

plt.plot(hidden_nodes, test_err[:, 0], label="relu激活(测试误差)", linestyle='-')
plt.plot(hidden_nodes, test_err[:, 1], label="logistic激活(测试误差)", linestyle='-.')
plt.plot(hidden_nodes, train_err[:, 0], label="relu激活(训练误差)", linestyle='-', linewidth=0.5)
plt.plot(hidden_nodes, train_err[:, 1], label="logistic激活(训练误差)", linestyle='-.', linewidth=0.5)
plt.title('隐藏节点数与误差')
plt.xlabel('隐藏节点数')
plt.ylabel('误差')
plt.xticks(hidden_nodes)
plt.legend()
plt.show()