神经网络分类

https://www.lanqiao.cn/courses/1029

对汽车燃料效率建模

数据集

http://labfile.oss.aliyuncs.com/courses/1029/mpg.csv

代码

# coding: utf-8
import matplotlib.pyplot as plt
import pandas as pd
from sklearn import model_selection, metrics
from sklearn import preprocessing
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
%matplotlib inline

# 读取数据集
df = pd.read_csv("mpg.csv", header=0)
# 将排量的参数转化为浮点类型的数字
df['displacement']=df['displacement'].astype(float)
# 从数据集中获取数据列，其中第一列（每加仑英里数）和最后一列（车名）将被忽略。
X = df[df.columns[1:8]]
y = df['mpg']

# 进行原始数据展示
plt.figure() # 创建一个新的图形
f, ax1 = plt.subplots()
for i in range (1,8):
    number = 420 + i
    ax1.locator_params(nbins=3)
    ax1 = plt.subplot(number)
    plt.title(list(df)[i])
    ax1.scatter(df[df.columns[i]],y) # 绘制数据点的散点图
plt.tight_layout(pad=0.4, w_pad=0.5, h_pad=1.0)

# 拆分数据集
X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y,test_size=0.25)
# 缩放数据以进行收敛性优化
scaler = preprocessing.StandardScaler()
# 设置变换参数
X_train = scaler.fit_transform(X_train)

# 建立一个含有两层全连接的DNN网络，其分别有十个和五个单元。
model = Sequential()
model.add(Dense(10, input_dim=7, kernel_initializer='normal', activation='relu'))
model.add(Dense(5, kernel_initializer='normal', activation='relu'))
model.add(Dense(1, kernel_initializer='normal'))

# 编译模型，将均方误差作为损失函数
model.compile(loss='mean_squared_error', optimizer='adam')
# 使用1000次训练匹配模型
model.fit(X_train, y_train, epochs=1000, validation_split=0.33, shuffle=True,verbose=2 )

# 准确度测试
score = metrics.mean_squared_error(model.predict(scaler.transform(X_test)),y_test)
print(" Total Mean Squared Error: " + str(score))

葡萄酒分类

数据集

http://labfile.oss.aliyuncs.com/courses/1029/wine.csv

代码

# coding: utf-8
import matplotlib.pyplot as plt
import pandas as pd
import tensorflow as tf
from sklearn.utils import shuffle
from sklearn import preprocessing
%matplotlib inline

df = pd.read_csv("wine.csv", header=0)
print (df.describe())

for i in range (1,8):
    number = 420 + i
    ax1 = plt.subplot(number)
    ax1.locator_params(nbins=3)
    plt.title
    ax1.scatter(df[df.columns[i]],df['Wine'])    # Plot a scatter draw of the  datapoints
plt.tight_layout(pad=0.4, w_pad=0.5, h_pad=1.0)

X = df[df.columns[1:13]].values
Y = df['Wine'].values-1
X, Y = shuffle (X, Y)

scaler = preprocessing.StandardScaler()
X = scaler.fit_transform(X)

class Linear(tf.keras.Model):
    def __init__(self):
        super().__init__()
        self.dense = tf.keras.layers.Dense(
            units=3,
            activation="softmax",
            kernel_initializer=tf.zeros_initializer(),
            bias_initializer=tf.zeros_initializer()
        )    # 全连接层 f.keras.layers.Dense 

    def call(self, input):
        output = self.dense(input)
        return output

tf.keras.backend.set_floatx('float64')
model = Linear()
optimizer = tf.keras.optimizers.SGD(learning_rate=0.01)    # 使用 SGD 优化器
sparse_categorical_accuracy = tf.keras.metrics.SparseCategoricalAccuracy()
for i in range(100):

    X,Y =shuffle (X, Y, random_state=1)
    Xtr=X[0:140,:]
    Ytr=Y[0:140]
    Xt=X[140:178,:]
    Yt=Y[140:178]
    Xtr, Ytr = shuffle (Xtr, Ytr, random_state=0)
    
    with tf.GradientTape() as tape:
        y_pred = model(Xtr)
        loss = tf.keras.losses.sparse_categorical_crossentropy(y_true=Ytr, y_pred=y_pred)    # 使用交叉熵损失
        loss = tf.reduce_mean(loss)
    grads = tape.gradient(loss, model.variables)
    optimizer.apply_gradients(grads_and_vars=zip(grads, model.variables))
    y_pred_test = model.predict(Xt)
    sparse_categorical_accuracy.update_state(y_true=Yt, y_pred=y_pred_test)    # 评估器，输出预测正确的样本数占总样本数的比例
    print("test accuracy: %f" % sparse_categorical_accuracy.result())