Keras学习（五）-泰坦尼克号幸存预测（全连接形式）

1.导入运行库

import numpy
import pandas as pd
from sklearn import preprocessing
numpy.random.seed(10)

2.数据准备

all_df = pd.read_excel("data/titanic3.xls")

#去除无用数据
cols=['survived','name','pclass' ,'sex', 'age', 'sibsp',
      'parch', 'fare', 'embarked']
all_df=all_df[cols]

#划分训练和测试集
msk = numpy.random.rand(len(all_df)) < 0.8
train_df = all_df[msk]
test_df = all_df[~msk]

#定义数据预处理函数
def PreprocessData(raw_df):
    df=raw_df.drop(['name'], axis=1)
    age_mean = df['age'].mean()
    df['age'] = df['age'].fillna(age_mean)
    fare_mean = df['fare'].mean()
    df['fare'] = df['fare'].fillna(fare_mean)
    df['sex']= df['sex'].map({'female':0, 'male': 1}).astype(int)
    x_OneHot_df = pd.get_dummies(data=df,columns=["embarked" ])

    ndarray = x_OneHot_df.values
    Features = ndarray[:,1:]
    Label = ndarray[:,0]

    minmax_scale = preprocessing.MinMaxScaler(feature_range=(0, 1))
    scaledFeatures=minmax_scale.fit_transform(Features)    
    
    return scaledFeatures,Label


#对已经划分好的数据集进行预处理
train_Features,train_Label=PreprocessData(train_df)
test_Features,test_Label=PreprocessData(test_df)

3.建立模型（全连接）

from keras.models import Sequential
from keras.layers import Dense,Dropout

#建立堆叠基础模型
model = Sequential()

#建立两层全连接，和输出两种情况的输出层
model.add(Dense(units=40, input_dim=9, 
                kernel_initializer='uniform', 
                activation='relu'))
model.add(Dense(units=30, 
                kernel_initializer='uniform', 
                activation='relu'))
model.add(Dense(units=1, 
                kernel_initializer='uniform',
                activation='sigmoid'))

4.训练模型

#定义模型格式
model.compile(loss='binary_crossentropy', 
              optimizer='adam', metrics=['accuracy'])

#导入训练数据
train_history =model.fit(x=train_Features, 
                         y=train_Label, 
                         validation_split=0.1, 
                         epochs=30, 
                         batch_size=30,verbose=1)
           

5.打印过程中的准确度和损失值

import matplotlib.pyplot as plt

#建立打印准确度和损失值的函数
def show_train_history(train_history,train,validation):
    plt.plot(train_history.history[train])
    plt.plot(train_history.history[validation])
    plt.title('Train History')
    plt.ylabel(train)
    plt.xlabel('Epoch')
    plt.legend(['train', 'validation'], loc='upper left')
    plt.show()

show_train_history(train_history,'acc','val_acc')
show_train_history(train_history,'loss','val_loss')

6.模型准确率计算

scores = model.evaluate(x=test_Features, 
                        y=test_Label)
print（scores[1]）              

7.加入主人公jack和rose的数据

Jack = pd.Series([0 ,'Jack',3, 'male'  , 23, 1, 0,  5.0000,'S'])
Rose = pd.Series([1 ,'Rose',1, 'female', 20, 1, 0, 100.0000,'S'])

JR_df = pd.DataFrame([list(Jack),list(Rose)],  
                  columns=['survived', 'name','pclass', 'sex', 
                   'age', 'sibsp','parch', 'fare','embarked'])

all_df=pd.concat([all_df,JR_df])

8.对所有数据进行预测

#对所有数据进行数据预处理
all_Features,Label=PreprocessData(all_df)

#进行预测
all_probability=model.predict(all_Features)

#在原有数据中添加生还几率的列
pd=all_df
pd.insert(len(all_df.columns),
          'probability',all_probability)

• 查看生还几率高缺没有生还的人

pd[(pd['survived']==0) &  (pd['probability']>0.9) ]