神经网络结果背后的直觉？

Question

我正在尝试建立一个神经网络来分类毒蘑菇，但结果不正确。该模型编译成功，但是有人可以提供直觉，说明为什么训练结果在几个 epoch 之后看起来如此准确。这似乎不正确，是数据预处理出错了吗？

数据集可以在这里找到：https://archive.ics.uci.edu/ml/machine-learning-databases/mushroom/agaricus-lepiota.data

代码如下：

import keras.utils
from sklearn import preprocessing
from sklearn.model_selection import train_test_split
import numpy as np
import pandas as pd

# seed weights
np.random.seed(3)

# import dataset
data = pd.read_csv('agaricus-lepiota.csv', delimiter=',')

# encode labels as integers so the can be one-hot-encoded which takes int matrix
le = preprocessing.LabelEncoder()
data = data.apply(le.fit_transform)

# one-hot-encode string data (now type int)
ohe = preprocessing.OneHotEncoder(sparse=False)
data = ohe.fit_transform(data)

X = data[:, 1:23]
Y = data[:, 0:1]

# split into test and train set
x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=.2, random_state=5)

# create model
model = Sequential()
model.add(Dense(500, input_dim=22, activation='relu'))
model.add(Dense(300, activation='relu'))
model.add(Dense(100, activation='relu'))
model.add(Dense(50, activation='relu'))
model.add(Dense(25, activation='relu'))
model.add(Dense(1, activation='sigmoid'))

model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])

model.fit(x_train, y_train, validation_data=(x_test, y_test), epochs=1000, batch_size=25)

Answer 1

一个时期就是很多次迭代（n=training_set_size/batch_size）。考虑到你有这么多层并且没有正则化，我会怀疑过度拟合。