看完之后,kaggle比赛的重点主要是:
- 对数据的认识
- 数据中的离群点\特殊点的分析和处理
- 重要的特征工程(feature engineering)
- 模型融合(model ensemble)
import pandas as pd #数据分析
import numpy as np #科学计算
from pandas import Series,DataFrame
data_train = pd.read_csv("E:\\kaggle\\titanic\\train.csv")
data_train
简单数据预处理
对缺失值的处理方法 :
- 缺失值的样本占的比例较高,舍弃,以免引入noise;
- 缺失值的样本适中,并且该属性是非连续值的特征属性,就将NAN作为一个新类别,加到该类别特征中;
- 缺失值的样本适中,并且该属性是连续值,给定一个步长STEP,将这个连续的属性离散化,然后再将NAN加入其中;
- 缺失值的数目不是很多。可以根据已有的值,拟合补上。
from sklearn.ensemble import RandomForestRegressor
### 使用 RandomForestClassifier 填补缺失的年龄属性
def set_missing_ages(df):
# 把已有的数值型特征取出来丢进Random Forest Regressor中
age_df = df[['Age','Fare', 'Parch', 'SibSp', 'Pclass']]
# 乘客分成已知年龄和未知年龄两部分
known_age = age_df[age_df.Age.notnull()].as_matrix()
unknown_age = age_df[age_df.Age.isnull()].as_matrix()
# y即目标年龄
y = known_age[:, 0]
# X即特征属性值
X = known_age[:, 1:]
# fit到RandomForestRegressor之中
rfr = RandomForestRegressor(random_state=0, n_estimators=2000, n_jobs=-1)
rfr.fit(X, y)
# 用得到的模型进行未知年龄结果预测
predictedAges = rfr.predict(unknown_age[:, 1::])
# 用得到的预测结果填补原缺失数据
df.loc[ (df.Age.isnull()), 'Age' ] = predictedAges
return df, rfr
def set_Cabin_type(df):
df.loc[ (df.Cabin.notnull()), 'Cabin' ] = "Yes"
df.loc[ (df.Cabin.isnull()), 'Cabin' ] = "No"
return df
data_train, rfr = set_missing_ages(data_train)
data_train = set_Cabin_type(data_train)
特征因子化,one_hot
dummies_Cabin = pd.get_dummies(data_train['Cabin'], prefix= 'Cabin')
dummies_Embarked = pd.get_dummies(data_train['Embarked'], prefix= 'Embarked')
dummies_Sex = pd.get_dummies(data_train['Sex'], prefix= 'Sex')
dummies_Pclass = pd.get_dummies(data_train['Pclass'], prefix= 'Pclass')
df = pd.concat([data_train, dummies_Cabin, dummies_Embarked, dummies_Sex, dummies_Pclass], axis=1)
df.drop(['Pclass', 'Name', 'Sex', 'Ticket', 'Cabin', 'Embarked'], axis=1, inplace=True)
df
一些属性的变化幅度大 ,对训练的收敛速度造成很大的影响,所以要将其进行区间缩放,即scaling,将它的变化范围控制到[-1,1]
import sklearn.preprocessing as preprocessing
from numpy import *
scaler = preprocessing.StandardScaler()
age_scale_param = scaler.fit(df['Age'].reshape(-1, 1))
df['Age_scaled'] = scaler.fit_transform(df['Age'].reshape(-1, 1), age_scale_param)
fare_scale_param = scaler.fit(df['Fare'].reshape(-1, 1))
df['Fare_scaled'] = scaler.fit_transform(df['Fare'].reshape(-1, 1), fare_scale_param)
df
逻辑回归模型
from sklearn import linear_model
# 用正则取出我们要的属性值
train_df = df.filter(regex='Survived|Age_.*|SibSp|Parch|Fare_.*|Cabin_.*|Embarked_.*|Sex_.*|Pclass_.*')
train_np = train_df.as_matrix()
# y即Survival结果
y = train_np[:, 0]
# X即特征属性值
X = train_np[:, 1:]
# fit到RandomForestRegressor之中
clf = linear_model.LogisticRegression(C=1.0, penalty='l1', tol=1e-6)
clf.fit(X, y)
clf
对test.csv重复上述操作
data_test = pd.read_csv("E:\\kaggle\\titanic\\test.csv")
data_test.loc[ (data_test.Fare.isnull()), 'Fare' ] = 0
# 接着我们对test_data做和train_data中一致的特征变换
# 首先用同样的RandomForestRegressor模型填上丢失的年龄
tmp_df = data_test[['Age','Fare', 'Parch', 'SibSp', 'Pclass']]
null_age = tmp_df[data_test.Age.isnull()].as_matrix()
# 根据特征属性X预测年龄并补上
X = null_age[:, 1:]
predictedAges = rfr.predict(X)
data_test.loc[ (data_test.Age.isnull()), 'Age' ] = predictedAges
data_test = set_Cabin_type(data_test)
dummies_Cabin = pd.get_dummies(data_test['Cabin'], prefix= 'Cabin')
dummies_Embarked = pd.get_dummies(data_test['Embarked'], prefix= 'Embarked')
dummies_Sex = pd.get_dummies(data_test['Sex'], prefix= 'Sex')
dummies_Pclass = pd.get_dummies(data_test['Pclass'], prefix= 'Pclass')
df_test = pd.concat([data_test, dummies_Cabin, dummies_Embarked, dummies_Sex, dummies_Pclass], axis=1)
df_test.drop(['Pclass', 'Name', 'Sex', 'Ticket', 'Cabin', 'Embarked'], axis=1, inplace=True)
df_test['Age_scaled'] = scaler.fit_transform(df_test['Age'].reshape(-1, 1), age_scale_param)
df_test['Fare_scaled'] = scaler.fit_transform(df_test['Fare'].reshape(-1, 1), fare_scale_param)
df_test
预测结果
est = df_test.filter(regex='Age_.*|SibSp|Parch|Fare_.*|Cabin_.*|Embarked_.*|Sex_.*|Pclass_.*')
predictions = clf.predict(test)
result = pd.DataFrame({'PassengerId':data_test['PassengerId'].as_matrix(), 'Survived':predictions.astype(np.int32)})
result.to_csv("E:\\kaggle\\titanic\\logistic_regression_predictions.csv", index=False)
把得到的model系数和feature关联起来看看
#各属性的相关性分析,如果为正,则为正相关
pd.DataFrame({"columns":list(train_df.columns)[1:], "coef":list(clf.coef_.T)})
# 分割数据,按照 训练数据:cv数据 = 7:3的比例
split_train, split_cv = cross_validation.train_test_split(df, test_size=0.3, random_state=0)
train_df = split_train.filter(regex='Survived|Age_.*|SibSp|Parch|Fare_.*|Cabin_.*|Embarked_.*|Sex_.*|Pclass_.*')
# 生成模型
clf = linear_model.LogisticRegression(C=1.0, penalty='l1', tol=1e-6)
clf.fit(train_df.as_matrix()[:,1:], train_df.as_matrix()[:,0])
# 对cross validation数据进行预测
cv_df = split_cv.filter(regex='Survived|Age_.*|SibSp|Parch|Fare_.*|Cabin_.*|Embarked_.*|Sex_.*|Pclass_.*')
predictions = clf.predict(cv_df.as_matrix()[:,1:])
origin_data_train = pd.read_csv("E:\\kaggle\\titanic\\train.csv")
bad_cases = origin_data_train.loc[origin_data_train['PassengerId'].isin(split_cv[predictions != cv_df.as_matrix()[:,0]]['PassengerId'].values)]
bad_cases
学习曲线
import numpy as np
import matplotlib.pyplot as plt
from sklearn.learning_curve import learning_curve
plt.rcParams['font.sans-serif']=['SimHei']
plt.rcParams['axes.unicode_minus']=False
# 用sklearn的learning_curve得到training_score和cv_score,使用matplotlib画出learning curve
def plot_learning_curve(estimator, title, X, y, ylim=None, cv=None, n_jobs=1,
train_sizes=np.linspace(.05, 1., 20), verbose=0, plot=True):
"""
画出data在某模型上的learning curve.
参数解释
----------
estimator : 你用的分类器。
title : 表格的标题。
X : 输入的feature,numpy类型
y : 输入的target vector
ylim : tuple格式的(ymin, ymax), 设定图像中纵坐标的最低点和最高点
cv : 做cross-validation的时候,数据分成的份数,其中一份作为cv集,其余n-1份作为training(默认为3份)
n_jobs : 并行的的任务数(默认1)
"""
train_sizes, train_scores, test_scores = learning_curve(
estimator, X, y, cv=cv, n_jobs=n_jobs, train_sizes=train_sizes, verbose=verbose)
train_scores_mean = np.mean(train_scores, axis=1)
train_scores_std = np.std(train_scores, axis=1)
test_scores_mean = np.mean(test_scores, axis=1)
test_scores_std = np.std(test_scores, axis=1)
if plot:
plt.figure(figsize=(12,8))
plt.title(title)
if ylim is not None:
plt.ylim(*ylim)
plt.xlabel(u"训练样本数")
plt.ylabel(u"得分")
plt.gca().invert_yaxis()
plt.grid()
plt.fill_between(train_sizes, train_scores_mean - train_scores_std, train_scores_mean + train_scores_std,
alpha=0.1, color="b")
plt.fill_between(train_sizes, test_scores_mean - test_scores_std, test_scores_mean + test_scores_std,
alpha=0.1, color="r")
plt.plot(train_sizes, train_scores_mean, 'o-', color="b", label=u"训练集上得分")
plt.plot(train_sizes, test_scores_mean, 'o-', color="r", label=u"交叉验证集上得分")
plt.legend(loc="best")
plt.draw()
plt.show()
plt.gca().invert_yaxis()
midpoint = ((train_scores_mean[-1] + train_scores_std[-1]) + (test_scores_mean[-1] - test_scores_std[-1])) / 2
diff = (train_scores_mean[-1] + train_scores_std[-1]) - (test_scores_mean[-1] - test_scores_std[-1])
return midpoint, diff
plot_learning_curve(clf, u"学习曲线", X, y)
模型融合
import numpy as np
import matplotlib.pyplot as plt
from sklearn.learning_curve import learning_curve
plt.rcParams['font.sans-serif']=['SimHei']
plt.rcParams['axes.unicode_minus']=False
# 用sklearn的learning_curve得到training_score和cv_score,使用matplotlib画出learning curve
def plot_learning_curve(estimator, title, X, y, ylim=None, cv=None, n_jobs=1,
train_sizes=np.linspace(.05, 1., 20), verbose=0, plot=True):
"""
画出data在某模型上的learning curve.
参数解释
----------
estimator : 你用的分类器。
title : 表格的标题。
X : 输入的feature,numpy类型
y : 输入的target vector
ylim : tuple格式的(ymin, ymax), 设定图像中纵坐标的最低点和最高点
cv : 做cross-validation的时候,数据分成的份数,其中一份作为cv集,其余n-1份作为training(默认为3份)
n_jobs : 并行的的任务数(默认1)
"""
train_sizes, train_scores, test_scores = learning_curve(
estimator, X, y, cv=cv, n_jobs=n_jobs, train_sizes=train_sizes, verbose=verbose)
train_scores_mean = np.mean(train_scores, axis=1)
train_scores_std = np.std(train_scores, axis=1)
test_scores_mean = np.mean(test_scores, axis=1)
test_scores_std = np.std(test_scores, axis=1)
if plot:
plt.figure(figsize=(12,8))
plt.title(title)
if ylim is not None:
plt.ylim(*ylim)
plt.xlabel(u"训练样本数")
plt.ylabel(u"得分")
plt.gca().invert_yaxis()
plt.grid()
plt.fill_between(train_sizes, train_scores_mean - train_scores_std, train_scores_mean + train_scores_std,
alpha=0.1, color="b")
plt.fill_between(train_sizes, test_scores_mean - test_scores_std, test_scores_mean + test_scores_std,
alpha=0.1, color="r")
plt.plot(train_sizes, train_scores_mean, 'o-', color="b", label=u"训练集上得分")
plt.plot(train_sizes, test_scores_mean, 'o-', color="r", label=u"交叉验证集上得分")
plt.legend(loc="best")
plt.draw()
plt.show()
plt.gca().invert_yaxis()
midpoint = ((train_scores_mean[-1] + train_scores_std[-1]) + (test_scores_mean[-1] - test_scores_std[-1])) / 2
diff = (train_scores_mean[-1] + train_scores_std[-1]) - (test_scores_mean[-1] - test_scores_std[-1])
return midpoint, diff
plot_learning_curve(clf, u"学习曲线", X, y)