用于多类分类的 ANN 模型

Question

我不知道问题是什么以及为什么会出现此错误：

ValueError: in user code:
ValueError: Shapes (None, 1) and (None, 6) are incompatible

谁能帮我处理这段代码？

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Activation,Dropout
from sklearn.preprocessing import MinMaxScaler
%matplotlib inline

df = pd.read_csv('test.csv')
dft = pd.read_csv('train.csv')

X_train = df.drop('label',axis=1).values
y_train = df['label'].values

X_test = dft.drop('label',axis=1).values
y_test = dft['label'].values

scaler = MinMaxScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)

model = Sequential()
model.add(Dense(units=30, activation='relu'))
model.add(Dense(units=15, activation='relu'))
model.add(Dense(6, activation='softmax'))

model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])
model.fit(x=X_train, y=y_train, epochs=15, batch_size=10, validation_data=(X_test, y_test))

Answer 1

问题是目标数组（y_train 和 y_test）的第二维长度等于 1，而模型预期为 6，假设输出层的神经元数量已设置等于 6。要解决此问题，您需要对目标进行一次热编码（您可以使用 scikit-learn OneHotEncoder）。如果您的目标确实有 6 个类，那么您的模型将起作用。

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from sklearn.datasets import make_classification
from sklearn.preprocessing import OneHotEncoder, MinMaxScaler
from sklearn.model_selection import train_test_split
tf.random.set_seed(0)

# generate the data
X, y = make_classification(n_classes=6, n_samples=1000, n_features=10, n_informative=10, n_redundant=0, random_state=42)
print(y.shape)
# (1000, )

# split the data
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)

# one-hot encode the target
enc = OneHotEncoder(sparse=False, handle_unknown='ignore')
enc.fit(y_train.reshape(-1, 1))
y_train = enc.transform(y_train.reshape(-1, 1))
y_test = enc.transform(y_test.reshape(-1, 1))
print(y_train.shape, y_test.shape)
# (750, 6) (250, 6)

# scale the features
scaler = MinMaxScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)

# define the model
model = Sequential()
model.add(Dense(units=30, activation='relu'))
model.add(Dense(units=15, activation='relu'))
model.add(Dense(6, activation='softmax'))

# fit the model
model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])
model.fit(x=X_train, y=y_train, epochs=3, batch_size=10, validation_data=(X_test, y_test))
# Epoch 1/3
# 75/75 [==============================] - 1s 2ms/step - loss: 1.7872 - accuracy: 0.2427 - val_loss: 1.7719 - val_accuracy: 0.2600
# Epoch 2/3
# 75/75 [==============================] - 0s 781us/step - loss: 1.7660 - accuracy: 0.2547 - val_loss: 1.7549 - val_accuracy: 0.2720
# Epoch 3/3
# 75/75 [==============================] - 0s 768us/step - loss: 1.7528 - accuracy: 0.2587 - val_loss: 1.7408 - val_accuracy: 0.3280