|
|
# -*- coding: utf-8 -*-
|
|
|
# ---
|
|
|
# jupyter:
|
|
|
# jupytext_format_version: '1.2'
|
|
|
# kernelspec:
|
|
|
# display_name: Python 3
|
|
|
# language: python
|
|
|
# name: python3
|
|
|
# language_info:
|
|
|
# codemirror_mode:
|
|
|
# name: ipython
|
|
|
# version: 3
|
|
|
# file_extension: .py
|
|
|
# mimetype: text/x-python
|
|
|
# name: python
|
|
|
# nbconvert_exporter: python
|
|
|
# pygments_lexer: ipython3
|
|
|
# version: 3.5.2
|
|
|
# ---
|
|
|
|
|
|
# # Exercise - 交通事故理赔审核预测
|
|
|
#
|
|
|
#
|
|
|
# 这个比赛的链接：http://sofasofa.io/competition.php?id=2
|
|
|
#
|
|
|
#
|
|
|
# * 任务类型：二元分类
|
|
|
#
|
|
|
# * 背景介绍：在交通摩擦（事故）发生后，理赔员会前往现场勘察、采集信息，这些信息往往影响着车主是否能够得到保险公司的理赔。训练集数据包括理赔人员在现场对该事故方采集的36条信息，信息已经被编码，以及该事故方最终是否获得理赔。我们的任务是根据这36条信息预测该事故方没有被理赔的概率。
|
|
|
#
|
|
|
# * 数据介绍：训练集中共有200000条样本，预测集中有80000条样本。
|
|
|
# 
|
|
|
#
|
|
|
# * 评价方法：Precision-Recall AUC
|
|
|
#
|
|
|
|
|
|
# ## Demo code
|
|
|
#
|
|
|
|
|
|
import pandas as pd
|
|
|
import numpy as np
|
|
|
import os
|
|
|
import matplotlib.pyplot as plt
|
|
|
# %matplotlib inline
|
|
|
|
|
|
# read data
|
|
|
homePath = "data"
|
|
|
trainPath = os.path.join(homePath, "train.csv")
|
|
|
testPath = os.path.join(homePath, "test.csv")
|
|
|
submitPath = os.path.join(homePath, "sample_submit.csv")
|
|
|
trainData = pd.read_csv(trainPath)
|
|
|
testData = pd.read_csv(testPath)
|
|
|
submitData = pd.read_csv(submitPath)
|
|
|
|
|
|
# 参照数据说明，CaseID这列是没有意义的编号，因此这里将他丢弃。
|
|
|
#
|
|
|
# ~drop()函数：axis指沿着哪个轴，0为行，1为列；inplace指是否在原数据上直接操作
|
|
|
#
|
|
|
|
|
|
# 去掉没有意义的一列
|
|
|
trainData.drop("CaseId", axis=1, inplace=True)
|
|
|
testData.drop("CaseId", axis=1, inplace=True)
|
|
|
|
|
|
# # 快速了解数据
|
|
|
#
|
|
|
# ~head()：默认显示前5行数据，可指定显示多行，例如.head(15)显示前15行
|
|
|
#
|
|
|
|
|
|
trainData.head(15)
|
|
|
|
|
|
|
|
|
# 显示数据简略信息，可以每列有多少非空的值，以及每列数据对应的数据类型。
|
|
|
#
|
|
|
#
|
|
|
|
|
|
trainData.info()
|
|
|
|
|
|
|
|
|
# ~hist():绘制直方图，参数figsize可指定输出图片的尺寸。
|
|
|
#
|
|
|
|
|
|
trainData.hist(figsize=(20, 20))
|
|
|
|
|
|
|
|
|
# 想要了解特征之间的相关性，可计算相关系数矩阵。然后可对某个特征来排序。
|
|
|
#
|
|
|
#
|
|
|
|
|
|
corr_matrix = trainData.corr()
|
|
|
corr_matrix["Evaluation"].sort_values(ascending=False) # ascending=False 降序排列
|
|
|
|
|
|
# 从训练集中分离标签
|
|
|
|
|
|
y = trainData['Evaluation']
|
|
|
trainData.drop("Evaluation", axis=1, inplace=True)
|
|
|
|
|
|
# 使用K-Means训练模型
|
|
|
#
|
|
|
# KMeans()：
|
|
|
# * `n_clusters`指要预测的有几个类；
|
|
|
# * `init`指初始化中心的方法，默认使用的是`k-means++`方法，而非经典的K-means方法的随机采样初始化，当然你可以设置为random使用随机初始化；
|
|
|
# * `n_jobs`指定使用CPU核心数，-1为使用全部CPU。
|
|
|
|
|
|
# +
|
|
|
# do k-means
|
|
|
from sklearn.cluster import KMeans
|
|
|
est = KMeans(n_clusters=2, init="k-means++", n_jobs=-1)
|
|
|
est.fit(trainData, y)
|
|
|
|
|
|
y_train = est.predict(trainData)
|
|
|
y_pred = est.predict(testData)
|
|
|
|
|
|
# 保存预测的结果
|
|
|
submitData['Evaluation'] = y_pred
|
|
|
submitData.to_csv("submit_data.csv", index=False)
|
|
|
|
|
|
# +
|
|
|
# calculate accuracy
|
|
|
from sklearn.metrics import accuracy_score
|
|
|
|
|
|
acc_train = accuracy_score(y, y_train)
|
|
|
print("acc_train = %f" % (acc_train))
|
|
|
# -
|
|
|
|
|
|
# ## 随机森林
|
|
|
#
|
|
|
# 使用K-means可能得到的结果没那么理想。在官网上，举办方给出了两个标杆模型，效果最好的是随机森林。以下是代码，读者可以自己测试。
|
|
|
#
|
|
|
#
|
|
|
|
|
|
# +
|
|
|
import pandas as pd
|
|
|
from sklearn.ensemble import RandomForestClassifier
|
|
|
from sklearn.metrics import accuracy_score
|
|
|
|
|
|
# 读取数据
|
|
|
train = pd.read_csv("data/train.csv")
|
|
|
test = pd.read_csv("data/test.csv")
|
|
|
submit = pd.read_csv("data/sample_submit.csv")
|
|
|
|
|
|
# 删除id
|
|
|
train.drop('CaseId', axis=1, inplace=True)
|
|
|
test.drop('CaseId', axis=1, inplace=True)
|
|
|
|
|
|
# 取出训练集的y
|
|
|
y_train = train.pop('Evaluation')
|
|
|
|
|
|
# 建立随机森林模型
|
|
|
clf = RandomForestClassifier(n_estimators=100, random_state=0)
|
|
|
clf.fit(train, y_train)
|
|
|
y_pred = clf.predict_proba(test)[:, 1]
|
|
|
|
|
|
# 输出预测结果至my_RF_prediction.csv
|
|
|
submit['Evaluation'] = y_pred
|
|
|
submit.to_csv('my_RF_prediction.csv', index=False)
|
|
|
|
|
|
|
|
|
|
|
|
# +
|
|
|
# freature importances
|
|
|
print(clf.feature_importances_)
|
|
|
|
|
|
# Train accuracy
|
|
|
from sklearn.metrics import accuracy_score
|
|
|
y_train_pred = clf.predict(train)
|
|
|
print(y_train_pred)
|
|
|
|
|
|
acc_train = accuracy_score(y_train, y_train_pred)
|
|
|
print("acc_train = %f" % (acc_train))
|
