资金流入流出预测-挑战Baseline (下）¶
特征提取¶
数据准备¶
In [1]:
import pandas as pd
import numpy as np
import datetime
import shap
import eli5
import seaborn as sns
import matplotlib.pyplot as plt
from mvtpy import mvtest
from wordcloud import WordCloud
from scipy import stats
from eli5.sklearn import PermutationImportance
from sklearn import tree
from sklearn.preprocessing import OneHotEncoder
from sklearn.linear_model import LinearRegression
from typing import *
import warnings
warnings.filterwarnings('ignore')
plt.rcParams['font.sans-serif']=['SimHei'] #用来正常显示中文标签
plt.rcParams['axes.unicode_minus']=False #用来正常显示负号
In [2]:
labels = ['total_purchase_amt','total_redeem_amt']
date_indexs = ['week','year','month','weekday','day']
In [3]:
# Load the balance data
def load_data(path: str = 'user_balance_table.csv')->pd.DataFrame:
data_balance = pd.read_csv(path)
return data_balance.reset_index(drop=True)
# add tiemstamp to dataset
def add_timestamp(data: pd.DataFrame, time_index: str = 'report_date')->pd.DataFrame:
data_balance = data.copy()
data_balance['date'] = pd.to_datetime(data_balance[time_index], format= "%Y%m%d")
data_balance['day'] = data_balance['date'].dt.day
data_balance['month'] = data_balance['date'].dt.month
data_balance['year'] = data_balance['date'].dt.year
data_balance['week'] = data_balance['date'].dt.week
data_balance['weekday'] = data_balance['date'].dt.weekday
return data_balance.reset_index(drop=True)
# total amount
def get_total_balance(data: pd.DataFrame, date: str = '2014-03-31')->pd.DataFrame:
df_tmp = data.copy()
df_tmp = df_tmp.groupby(['date'])['total_purchase_amt','total_redeem_amt'].sum()
df_tmp.reset_index(inplace=True)
return df_tmp[(df_tmp['date']>= date)].reset_index(drop=True)
# Generate the test data
def generate_test_data(data: pd.DataFrame)->pd.DataFrame:
total_balance = data.copy()
start = datetime.datetime(2014,9,1)
testdata = []
while start != datetime.datetime(2014,10,15):
temp = [start, np.nan, np.nan]
testdata.append(temp)
start += datetime.timedelta(days = 1)
testdata = pd.DataFrame(testdata)
testdata.columns = total_balance.columns
total_balance = pd.concat([total_balance, testdata], axis = 0)
total_balance = total_balance.reset_index(drop=True)
return total_balance.reset_index(drop=True)
# Load user's information
def load_user_information(path: str = 'user_profile_table.csv')->pd.DataFrame:
return pd.read_csv(path)
In [4]:
# 读取数据集
balance_data = load_data('Dataset/user_balance_table.csv')
balance_data = add_timestamp(balance_data, time_index='report_date')
total_balance = get_total_balance(balance_data)
total_balance = generate_test_data(total_balance)
total_balance = add_timestamp(total_balance, 'date')
user_information = load_user_information('Dataset/user_profile_table.csv')
1. 基于日期的静态特征¶
1.1 提取 is 特征¶
In [5]:
# 获取节假日集合
def get_holiday_set()->Set[datetime.date]:
holiday_set = set()
# 清明节
holiday_set = holiday_set | {datetime.date(2014,4,5), datetime.date(2014,4,6), datetime.date(2014,4,7)}
# 劳动节
holiday_set = holiday_set | {datetime.date(2014,5,1), datetime.date(2014,5,2), datetime.date(2014,5,3)}
# 端午节
holiday_set = holiday_set | {datetime.date(2014,5,31), datetime.date(2014,6,1), datetime.date(2014,6,2)}
# 中秋节
holiday_set = holiday_set | {datetime.date(2014,9,6), datetime.date(2014,9,7), datetime.date(2014,9,8)}
# 国庆节
holiday_set = holiday_set | {datetime.date(2014,10,1), datetime.date(2014,10,2), datetime.date(2014,10,3),\
datetime.date(2014,10,4), datetime.date(2014,10,5), datetime.date(2014,10,6),\
datetime.date(2014,10,7)}
# 中秋节
holiday_set = holiday_set | {datetime.date(2013,9,19), datetime.date(2013,9,20), datetime.date(2013,9,21)}
# 国庆节
holiday_set = holiday_set | {datetime.date(2013,10,1), datetime.date(2013,10,2), datetime.date(2013,10,3),\
datetime.date(2013,10,4), datetime.date(2013,10,5), datetime.date(2013,10,6),\
datetime.date(2013,10,7)}
return holiday_set
In [6]:
# 提取所有 is特征
def extract_is_feature(data: pd.DataFrame)->pd.DataFrame:
total_balance = data.copy().reset_index(drop=True)
# 是否是Weekend
total_balance['is_weekend'] = 0
total_balance.loc[total_balance['weekday'].isin((5,6)), 'is_weekend'] = 1
# 是否是假期
total_balance['is_holiday'] = 0
total_balance.loc[total_balance['date'].isin(get_holiday_set()), 'is_holiday'] = 1
# 是否是节假日的第一天
last_day_flag = 0
total_balance['is_firstday_of_holiday'] = 0
for index, row in total_balance.iterrows():
if last_day_flag == 0 and row['is_holiday'] == 1:
total_balance.loc[index, 'is_firstday_of_holiday'] = 1
last_day_flag = row['is_holiday']
# 是否是节假日的最后一天
total_balance['is_lastday_of_holiday'] = 0
for index, row in total_balance.iterrows():
if row['is_holiday'] == 1 and total_balance.loc[index+1, 'is_holiday'] == 0:
total_balance.loc[index, 'is_lastday_of_holiday'] = 1
# 是否是节假日后的上班第一天
total_balance['is_firstday_of_work'] = 0
last_day_flag = 0
for index, row in total_balance.iterrows():
if last_day_flag == 1 and row['is_holiday'] == 0:
total_balance.loc[index, 'is_firstday_of_work'] = 1
last_day_flag = row['is_lastday_of_holiday']
# 是否不用上班
total_balance['is_work'] = 1
total_balance.loc[(total_balance['is_holiday'] == 1) | (total_balance['is_weekend'] == 1), 'is_work'] = 0
special_work_day_set = {datetime.date(2014,5,4), datetime.date(2014,9,28)}
total_balance.loc[total_balance['date'].isin(special_work_day_set), 'is_work'] = 1
# 是否明天要上班
total_balance['is_gonna_work_tomorrow'] = 0
for index, row in total_balance.iterrows():
if index == len(total_balance)-1:
break
if row['is_work'] == 0 and total_balance.loc[index+1, 'is_work'] == 1:
total_balance.loc[index, 'is_gonna_work_tomorrow'] = 1
# 昨天上班了吗
total_balance['is_worked_yestday'] = 0
for index, row in total_balance.iterrows():
if index <= 1:
continue
if total_balance.loc[index-1, 'is_work'] == 1:
total_balance.loc[index, 'is_worked_yestday'] = 1
# 是否是放假前一天
total_balance['is_lastday_of_workday'] = 0
for index, row in total_balance.iterrows():
if index == len(total_balance)-1:
break
if row['is_holiday'] == 0 and total_balance.loc[index+1, 'is_holiday'] == 1:
total_balance.loc[index, 'is_lastday_of_workday'] = 1
# 是否周日要上班
total_balance['is_work_on_sunday'] = 0
for index, row in total_balance.iterrows():
if index == len(total_balance)-1:
break
if row['weekday'] == 6 and row['is_work'] == 1:
total_balance.loc[index, 'is_work_on_sunday'] = 1
# 是否是月初第一天
total_balance['is_firstday_of_month'] = 0
total_balance.loc[total_balance['day'] == 1, 'is_firstday_of_month'] = 1
# 是否是月初第二天
total_balance['is_secday_of_month'] = 0
total_balance.loc[total_balance['day'] == 2, 'is_secday_of_month'] = 1
# 是否是月初
total_balance['is_premonth'] = 0
total_balance.loc[total_balance['day'] <= 10, 'is_premonth'] = 1
# 是否是月中
total_balance['is_midmonth'] = 0
total_balance.loc[(10 < total_balance['day']) & (total_balance['day'] <= 20), 'is_midmonth'] = 1
# 是否是月末
total_balance['is_tailmonth'] = 0
total_balance.loc[20 < total_balance['day'], 'is_tailmonth'] = 1
# 是否是每个月第一个周
total_balance['is_first_week'] = 0
total_balance.loc[total_balance['week'] % 4 == 1, 'is_first_week'] = 1
# 是否是每个月第一个周
total_balance['is_second_week'] = 0
total_balance.loc[total_balance['week'] % 4 == 2, 'is_second_week'] = 1
# 是否是每个月第一个周
total_balance['is_third_week'] = 0
total_balance.loc[total_balance['week'] % 4 == 3, 'is_third_week'] = 1
# 是否是每个月第四个周
total_balance['is_fourth_week'] = 0
total_balance.loc[total_balance['week'] % 4 == 0, 'is_fourth_week'] = 1
return total_balance.reset_index(drop=True)
In [7]:
# 提取is特征到数据集
total_balance = extract_is_feature(total_balance)
In [8]:
# 编码翌日特征
def encode_data(data: pd.DataFrame, feature_name:str = 'weekday', encoder=OneHotEncoder())->pd.DataFrame():
total_balance = data.copy()
week_feature = encoder.fit_transform(np.array(total_balance[feature_name]).reshape(-1, 1)).toarray()
week_feature = pd.DataFrame(week_feature,columns= [feature_name + '_onehot_'+ str(x) for x in range(len(week_feature[0]))])
#featureWeekday = pd.concat([total_balance, week_feature], axis = 1).drop(feature_name, axis=1)
featureWeekday = pd.concat([total_balance, week_feature], axis = 1)
return featureWeekday
In [9]:
# 编码翌日特征到数据集
total_balance = encode_data(total_balance)
In [10]:
# 生成is特征集合
feature = total_balance[[x for x in total_balance.columns if x not in date_indexs]]
In [11]:
feature.head()
Out[11]:
1.2 is特征的下标签分布分析¶
In [12]:
# 绘制箱型图
def draw_boxplot(data: pd.DataFrame)->None:
f, axes = plt.subplots(7, 4, figsize=(18, 24))
global date_indexs, labels
count = 0
for i in [x for x in data.columns if x not in date_indexs + labels + ['date']]:
sns.boxenplot(x=i, y='total_purchase_amt', data=data, ax=axes[count // 4][count % 4])
count += 1
In [13]:
draw_boxplot(feature)
In [14]:
## 剔除看起来较差的特征
purchase_feature_seems_useless = [
#样本量太少，建模时无效；但若确定这是一个有用规则，可以对结果做修正
'is_work_on_sunday',
#中位数差异不明显
'is_first_week'
]
1.3 IS 特征的相关性分析¶
In [15]:
# 画相关性热力图
def draw_correlation_heatmap(data: pd.DataFrame, way:str = 'pearson')->None:
feature = data.copy()
plt.figure(figsize=(20,10))
plt.title('The ' + way +' coleration between total purchase and each feature')
sns.heatmap(feature[[x for x in feature.columns if x not in ['total_redeem_amt', 'date'] ]].corr(way),linecolor='white',
linewidths=0.1,
cmap="RdBu")
In [16]:
draw_correlation_heatmap(feature, 'spearman')
In [17]:
# 剔除相关性较低的特征
temp = np.abs(feature[[x for x in feature.columns
if x not in ['total_redeem_amt', 'date'] ]].corr('spearman')['total_purchase_amt'])
feature_low_correlation = list(set(temp[temp < 0.1].index))
In [18]:
feature_low_correlation
Out[18]:
2. 基于距离的特征¶
2.1 距离特征提取¶
In [19]:
# 提取距离特征
def extract_distance_feature(data: pd.DataFrame)->pd.DataFrame:
total_balance = data.copy()
# 距离放假还有多少天
total_balance['dis_to_nowork'] = 0
for index, row in total_balance.iterrows():
if row['is_work'] == 0:
step = 1
flag = 1
while flag:
if index - step >= 0 and total_balance.loc[index - step, 'is_work'] == 1:
total_balance.loc[index - step, 'dis_to_nowork'] = step
step += 1
else:
flag = 0
total_balance['dis_from_nowork'] = 0
step = 0
for index, row in total_balance.iterrows():
step += 1
if row['is_work'] == 1:
total_balance.loc[index, 'dis_from_nowork'] = step
else:
step = 0
# 距离上班还有多少天
total_balance['dis_to_work'] = 0
for index, row in total_balance.iterrows():
if row['is_work'] == 1:
step = 1
flag = 1
while flag:
if index - step >= 0 and total_balance.loc[index - step, 'is_work'] == 0:
total_balance.loc[index - step, 'dis_to_work'] = step
step += 1
else:
flag = 0
total_balance['dis_from_work'] = 0
step = 0
for index, row in total_balance.iterrows():
step += 1
if row['is_work'] == 0:
total_balance.loc[index, 'dis_from_work'] = step
else:
step = 0
# 距离节假日还有多少天
total_balance['dis_to_holiday'] = 0
for index, row in total_balance.iterrows():
if row['is_holiday'] == 1:
step = 1
flag = 1
while flag:
if index - step >= 0 and total_balance.loc[index - step, 'is_holiday'] == 0:
total_balance.loc[index - step, 'dis_to_holiday'] = step
step += 1
else:
flag = 0
total_balance['dis_from_holiday'] = 0
step = 0
for index, row in total_balance.iterrows():
step += 1
if row['is_holiday'] == 0:
total_balance.loc[index, 'dis_from_holiday'] = step
else:
step = 0
# 距离节假日最后一天还有多少天
total_balance['dis_to_holiendday'] = 0
for index, row in total_balance.iterrows():
if row['is_lastday_of_holiday'] == 1:
step = 1
flag = 1
while flag:
if index - step >= 0 and total_balance.loc[index - step, 'is_lastday_of_holiday'] == 0:
total_balance.loc[index - step, 'dis_to_holiendday'] = step
step += 1
else:
flag = 0
total_balance['dis_from_holiendday'] = 0
step = 0
for index, row in total_balance.iterrows():
step += 1
if row['is_lastday_of_holiday'] == 0:
total_balance.loc[index, 'dis_from_holiendday'] = step
else:
step = 0
# 距离月初第几天
total_balance['dis_from_startofmonth'] = np.abs(total_balance['day'])
# 距离月的中心点有几天
total_balance['dis_from_middleofmonth'] = np.abs(total_balance['day'] - 15)
# 距离星期的中心有几天
total_balance['dis_from_middleofweek'] = np.abs(total_balance['weekday'] - 3)
# 距离星期日有几天
total_balance['dis_from_endofweek'] = np.abs(total_balance['weekday'] - 6)
return total_balance
In [20]:
# 拼接距离特征到原数据集
total_balance = extract_distance_feature(total_balance)
In [21]:
total_balance.shape
Out[21]:
2.2 距离特征分析¶
In [22]:
# 获取距离特征的列名
feature = total_balance[[x for x in total_balance.columns if x not in date_indexs]]
dis_feature_indexs = [x for x in feature.columns if (x not in date_indexs + labels + ['date']) & ('dis' in x)]
In [23]:
# 画点线
def draw_point_feature(data: pd.DataFrame)->None:
feature = data.copy()
f, axes = plt.subplots(data.shape[1] // 3, 3, figsize=(30, data.shape[1] // 3 * 4))
count = 0
for i in [x for x in feature.columns if (x not in date_indexs + labels + ['date'])]:
sns.pointplot(x=i, y="total_purchase_amt",
markers=["^", "o"], linestyles=["-", "--"],
kind="point", data=feature, ax=axes[count // 3][count % 3] if data.shape[1] > 3 else axes[count])
count += 1
In [24]:
draw_point_feature(feature[['total_purchase_amt'] + dis_feature_indexs])
In [25]:
# 处理距离过远的时间点
def dis_change(x):
if x > 5:
x = 10
return x
In [26]:
# 处理特殊距离
dis_holiday_feature = [x for x in total_balance.columns if 'dis' in x and 'holi' in x]
dis_month_feature = [x for x in total_balance.columns if 'dis' in x and 'month' in x]
total_balance[dis_holiday_feature] = total_balance[dis_holiday_feature].applymap(dis_change)
total_balance[dis_month_feature] = total_balance[dis_month_feature].applymap(dis_change)
In [27]:
feature = total_balance[[x for x in total_balance.columns if x not in date_indexs]]
In [28]:
# 画处理后的点线图
draw_point_feature(feature[['total_purchase_amt'] + dis_feature_indexs])
In [29]:
## 剔除看起来用处不大的特征
purchase_feature_seems_useless += [
#即使做了处理，但方差太大，不可信，规律不明显
'dis_to_holiday',
#方差太大，不可信
'dis_from_startofmonth',
#方差太大，不可信
'dis_from_middleofmonth'
]
In [30]:
# 画出相关性图
draw_correlation_heatmap(feature[['total_purchase_amt'] + dis_feature_indexs])
In [31]:
# 剔除相关性较差的特征
temp = np.abs(feature[[x for x in feature.columns
if ('dis' in x) | (x in ['total_purchase_amt']) ]].corr()['total_purchase_amt'])
feature_low_correlation += list(set(temp[temp < 0.1].index) )
In [32]:
feature_low_correlation
Out[32]:
3. 波峰波谷特征¶
3.1 提取波峰特征¶
In [33]:
# 观察波峰特点
fig = plt.figure(figsize=(15,12))
for i in range(6, 10):
plt.subplot(5,1,i - 5)
total_balance_2 = total_balance[(total_balance['date'] >= datetime.datetime(2014,8,1)) & (total_balance['date'] < datetime.datetime(2014,9,1))]
sns.pointplot(x=total_balance_2['day'],y=total_balance_2['total_purchase_amt'])
plt.legend().set_title('Month:' + str(i))
In [34]:
#Purchase
#0401(周二) 0406（周日，清明节第二天）
#0410(周四，与周二近似) 0412（周六，与周日近似）
#0415（周二） 0420（周日）
#0424（周四，与周二在近似水平） 0427（周日）
#0429（周二） 0502（周五，劳动节第二天）
#0507（周三，与周二差异较大，可能受劳务节影响） 0511（周日）
#0512（周一，与周二有一定差距） 0518（周日）
#0519（周二） 0525（周日）
#0526（周一，与周二有一定差距） 0531（周六，月末）
#0605（周四，与周二差异大，可能受端午节影响） 0607（周六，可能受端午节影响）
#0609（周一，与周二近似） 0615（周日）
#0616（周一，与周二差异大） 0622（周日）
#0626（周四，与周二差异不大） 0629（周日）
#0701（周二） 0705（周六，与周日差距不大）
#0707（周一，与周二有差距） 0713（周日）
#0716（周三，与周二有一定差距） 0720（周日）
#0721（周一，与周二有明显差距） 0726（周六，与周日近似）
#0728（周一，与周二有明显差距） 0803（周日）
#0805（周二） 0809（周六，与周日有较大差距）
#0811（周一，有周二有较大差距） 0817（周日）
#0818（周一，与周二差距不大） 0824（周日）
In [35]:
# 设定波峰日期
def extract_peak_feature(data: pd.DataFrame)->pd.DataFrame:
total_balance = data.copy()
# 距离purchase波峰（即周二）有几天
total_balance['dis_from_purchase_peak'] = np.abs(total_balance['weekday'] - 1)
# 距离purchase波谷（即周日）有几天，与dis_from_endofweek相同
total_balance['dis_from_purchase_valley'] = np.abs(total_balance['weekday'] - 6)
return total_balance
In [36]:
# 提取波峰特征
total_balance = extract_peak_feature(total_balance)
feature = total_balance[[x for x in total_balance.columns if x not in date_indexs]]
In [37]:
feature.head()
Out[37]:
3.2 分析波峰特征¶
In [38]:
draw_point_feature(feature[['total_purchase_amt'] + ['dis_from_purchase_peak','dis_from_purchase_valley']])
3.3 分析波峰特征相关性¶
In [39]:
temp = np.abs(feature[[x for x in feature.columns if ('peak' in x) or ('valley' in x) or (x in ['total_purchase_amt']) ]].corr()['total_purchase_amt'])
temp
Out[39]:
4. 周期因子分析¶
4.1 提取周期因子¶
In [40]:
def generate_rate(df, month_index):
total_balance = df.copy()
pure_balance = total_balance[['date','total_purchase_amt','total_redeem_amt']]
pure_balance = pure_balance[(pure_balance['date'] >= datetime.datetime(2014,3,1)) & (pure_balance['date'] < datetime.datetime(2014, month_index, 1))]
pure_balance['weekday'] = pure_balance['date'].dt.weekday
pure_balance['day'] = pure_balance['date'].dt.day
pure_balance['week'] = pure_balance['date'].dt.week
pure_balance['month'] = pure_balance['date'].dt.month
weekday_rate = pure_balance[['weekday']+labels].groupby('weekday',as_index=False).mean()
for name in labels:
weekday_rate = weekday_rate.rename(columns={name: name+'_weekdaymean'})
weekday_rate['total_purchase_amt_weekdaymean'] /= np.mean(pure_balance['total_purchase_amt'])
weekday_rate['total_redeem_amt_weekdaymean'] /= np.mean(pure_balance['total_redeem_amt'])
pure_balance = pd.merge(pure_balance, weekday_rate, on='weekday', how='left')
weekday_count = pure_balance[['day','weekday','date']].groupby(['day','weekday'],as_index=False).count()
weekday_count = pd.merge(weekday_count, weekday_rate, on = 'weekday')
weekday_count['total_purchase_amt_weekdaymean'] *= weekday_count['date'] / (len(set(pure_balance['month'])) - 1)
weekday_count['total_redeem_amt_weekdaymean'] *= weekday_count['date'] / (len(set(pure_balance['month'])) - 1)
day_rate = weekday_count.drop(['weekday','date'],axis=1).groupby('day',as_index=False).sum()
weekday_rate.columns = ['weekday','purchase_weekdayrate','redeem_weekdayrate']
day_rate.columns = ['day','purchase_dayrate','redeem_dayrate']
day_rate['date'] = datetime.datetime(2014, month_index, 1)
for index, row in day_rate.iterrows():
if month_index in (2,4,6,9) and row['day'] == 31:
continue
day_rate.loc[index, 'date'] = datetime.datetime(2014, month_index, int(row['day']))
day_rate['weekday'] = day_rate['date'].dt.weekday
day_rate = pd.merge(day_rate, weekday_rate, on='weekday')
day_rate['purchase_dayrate'] = day_rate['purchase_weekdayrate'] / day_rate['purchase_dayrate']
day_rate['redeem_dayrate'] = day_rate['redeem_weekdayrate'] / day_rate['redeem_dayrate']
weekday_rate['month'] = month_index
day_rate['month'] = month_index
return weekday_rate, day_rate[['day','purchase_dayrate','redeem_dayrate','month']].sort_values('day')
In [41]:
# 生成周期因子并合并到数据集
weekday_rate_list = []
day_rate_list = []
for i in range(3, 10):
weekday_rate, day_rate = generate_rate(total_balance, i)
weekday_rate_list.append(weekday_rate.reset_index(drop=True))
day_rate_list.append(day_rate.reset_index(drop=True))
weekday_rate_list = pd.concat(weekday_rate_list).reset_index(drop=True)
day_rate_list = pd.concat(day_rate_list).reset_index(drop=True)
total_balance = pd.merge(total_balance, weekday_rate_list, on=['weekday','month'], how='left')
total_balance = pd.merge(total_balance, day_rate_list, on=['day','month'], how='left')
In [42]:
# 对周期因子进行特殊处理
for i in [x for x in total_balance.columns
if 'rate' in x and x not in labels + date_indexs]:
total_balance[i] = total_balance[i].fillna(np.nanmedian(total_balance[i]))
In [43]:
total_balance.head()
Out[43]:
4.2 分析周期因子的相关性¶
In [44]:
# 画出相关性图
draw_correlation_heatmap(total_balance[['total_purchase_amt']
+ [x for x in total_balance.columns
if 'rate' in x and x not in labels + date_indexs]])
In [45]:
# 剔除相关性低的特征
feature = total_balance.drop(date_indexs, axis=1)
feature
Out[45]:
5. 动态时序分析¶
5.1 提取动态特征¶
In [46]:
## 提取动态特征
def get_amtfeature_with_time(data: pd.DataFrame)->pd.DataFrame:
df_tmp_ = data[labels + date_indexs + ['date']].copy()
total_balance = data.copy()
df_tmp_ = df_tmp_[(df_tmp_['date']>=datetime.datetime(2014,3,3))]
df_tmp_['weekday'] = df_tmp_['date'].dt.weekday + 1
df_tmp_['week'] = df_tmp_['date'].dt.week - min(df_tmp_['date'].dt.week) + 1
df_tmp_['day'] = df_tmp_['date'].dt.day
df_tmp_['month'] = df_tmp_['date'].dt.month
df_tmp_.reset_index(inplace=True)
del df_tmp_['index']
df_purchase = pd.DataFrame(columns = ['weekday1','weekday2','weekday3','weekday4',
'weekday5','weekday6','weekday7'])
count = 0
for i in range(len(df_tmp_)):
df_purchase.loc[count,'weekday'+str(df_tmp_.loc[i,'weekday'])] = df_tmp_.loc[i,'total_purchase_amt']
if df_tmp_.loc[i,'weekday'] == 7:
count = count + 1
df_tmp_['purchase_weekday_median'] = np.nan
df_tmp_['purchase_weekday_mean'] = np.nan
df_tmp_['purchase_weekday_min'] = np.nan
df_tmp_['purchase_weekday_max'] = np.nan
df_tmp_['purchase_weekday_std'] = np.nan
df_tmp_['purchase_weekday_skew'] = np.nan
for i in range(len(df_tmp_)):
#从2014年3月31日开始统计
if i > 4*7-1:
df_tmp_.loc[i,'purchase_weekday_median'] = df_purchase.loc[:df_tmp_.loc[i,'week']-2,
'weekday'+str(df_tmp_.loc[i,'weekday'])].median()
df_tmp_.loc[i,'purchase_weekday_mean'] = df_purchase.loc[:df_tmp_.loc[i,'week']-2,
'weekday'+str(df_tmp_.loc[i,'weekday'])].mean()
df_tmp_.loc[i,'purchase_weekday_min'] = df_purchase.loc[:df_tmp_.loc[i,'week']-2,
'weekday'+str(df_tmp_.loc[i,'weekday'])].min()
df_tmp_.loc[i,'purchase_weekday_max'] = df_purchase.loc[:df_tmp_.loc[i,'week']-2,
'weekday'+str(df_tmp_.loc[i,'weekday'])].max()
df_tmp_.loc[i,'purchase_weekday_std'] = df_purchase.loc[:df_tmp_.loc[i,'week']-2,
'weekday'+str(df_tmp_.loc[i,'weekday'])].std()
df_tmp_.loc[i,'purchase_weekday_skew'] = df_purchase.loc[:df_tmp_.loc[i,'week']-2,
'weekday'+str(df_tmp_.loc[i,'weekday'])].skew()
colList = ['purchase_weekday_median','purchase_weekday_mean','purchase_weekday_min',
'purchase_weekday_max','purchase_weekday_std','purchase_weekday_skew']
total_balance = pd.merge(total_balance, df_tmp_[colList+['day','month']], on=['day','month'], how='left')
return total_balance
In [47]:
# 合并特征到数据集
total_balance = get_amtfeature_with_time(total_balance)
In [48]:
# 对动态特征做特殊处理
for i in [x for x in total_balance.columns
if '_weekday_' in x and x not in labels + date_indexs]:
total_balance[i] = total_balance[i].fillna(np.nanmedian(total_balance[i]))
In [49]:
total_balance.head()
Out[49]:
5.2 分析动态特征相关性¶
In [50]:
# 绘制动态特征的相关性图
draw_correlation_heatmap(total_balance[['total_purchase_amt'] +
['purchase_weekday_median',
'purchase_weekday_mean','purchase_weekday_min',
'purchase_weekday_max','purchase_weekday_std',
'purchase_weekday_skew'
]])
In [51]:
feature[labels + ['dis_to_nowork', 'dis_to_work', 'dis_from_work', 'purchase_weekdayrate',
'redeem_dayrate', 'weekday_onehot_5', 'weekday_onehot_6',
'dis_from_nowork', 'is_holiday', 'weekday_onehot_1', 'weekday_onehot_2',
'weekday_onehot_0', 'dis_from_middleofweek', 'dis_from_holiendday',
'weekday_onehot_3', 'is_lastday_of_holiday', 'is_firstday_of_holiday',
'weekday_onehot_4', 'is_worked_yestday', 'is_second_week',
'is_third_week', 'dis_from_startofmonth', 'dis_from_holiday', 'total_purchase_amt',
'total_redeem_amt', 'date']].to_csv('Feature/0615_residual_purchase_origined.csv', index=False)
特征劣汰剔除¶
2.1 剔除无法有效分割数据集的特征¶
In [52]:
# 画出各个特征分割数据集的分布估计图
plt.figure(figsize=(4 * 6, 6 * len(feature.columns) / 6))
count = 0
for i in [x for x in feature.columns
if (x not in labels + date_indexs + ['date'])
& ('amt' not in x) & ('dis' not in x) & ('rate' not in x)]:
count += 1
if feature[feature[i] == 0].empty:
continue
plt.subplot(len(feature.columns) / 4, 4, count)
ax = sns.kdeplot(feature[feature[i] == 0]['total_purchase_amt'], label= str(i) + ' = 0, 买')
ax = sns.kdeplot(feature[feature[i] == 1]['total_purchase_amt'], label= str(i) + ' = 1, 买')
plt.rcParams.update({'font.size': 8})
plt.legend(loc = 0)
plt.title(str(i))
In [53]:
# 剔除对数据集划分不明显的特征
purchase_feature_seems_useless += ['is_gonna_work_tomorrow','is_fourth_week','weekday_onehot_4']
2.2 使用MVTest挽回一些有依赖性但是不相关的特征¶
In [54]:
feature_low_correlation
Out[54]:
In [55]:
# MVtest Ref: https://github.com/ChuanyuXue/MVTest
l = mvtest.mvtest()
name_list = []
Tn_list = []
p_list = []
for i in [i for i in feature_low_correlation if 'is' in i or 'discret' in i]:
pair = l.test(feature['total_purchase_amt'], feature[i])
name_list.append(str(i))
Tn_list.append(pair['Tn'])
p_list.append(pair['p-value'][0])
temp = pd.DataFrame([name_list,Tn_list]).T.sort_values(1)
temp[1] = np.abs(temp[1])
feature_saved_from_mv_purchase = list(temp.sort_values(1, ascending=False)[temp[1] > 0.5984][0])
2.3 剔除复共线特征¶
In [56]:
feature = feature[[x for x in feature.columns
if (x not in feature_low_correlation + purchase_feature_seems_useless) or\
(x in feature_saved_from_mv_purchase )]]
In [57]:
purchase_cors = feature.corr()
purchase_cors['total_purchase_amt'] = np.abs(purchase_cors['total_purchase_amt'])
feature_lists = list(purchase_cors.sort_values(by='total_purchase_amt',ascending=False).index)[2:]
feature_temp = feature.dropna()
In [58]:
# 这里要注意 保留的时候按照相关性降序排序 剔除按照相关性升序排序的顺序
thershold = 0.8
for i in range(len(feature_lists)):
for k in range(len(feature_lists)-1, -1, -1):
if i >= len(feature_lists) or k >= len(feature_lists) or i == k:
break
if np.abs(np.corrcoef(feature_temp[feature_lists[i]], feature_temp[feature_lists[k]])[0][1]) > thershold:
higher_feature_temp = feature_temp[feature_lists[i]] * feature_temp[feature_lists[k]]
if np.abs(np.corrcoef(feature_temp[feature_lists[i]], higher_feature_temp)[0][1]) <= thershold:
name = str(feature_lists[i]) + '%%%%' + str(feature_lists[k])
feature_temp[name] = higher_feature_temp
feature[name] = feature[feature_lists[i]] * feature[feature_lists[k]]
feature_lists.append(name)
feature_temp = feature_temp.drop(feature_lists[k], axis=1)
feature_lists.remove(feature_lists[k])
In [59]:
feature = feature[[x for x in feature_lists if x not in labels] + labels + ['date']]
In [60]:
feature_lists
Out[60]:
In [61]:
feature.to_csv('Feature/purchase_feature_droped_0614.csv',index=False)
选出优胜特征¶
In [62]:
# 分割数据集
def split_data_underline(data):
trainset = data[(datetime.datetime(2014,4,1) <= data['date']) & (data['date'] < datetime.datetime(2014,8,1))]
testset = data[(datetime.datetime(2014,8,1) <= data['date']) & (data['date'] < datetime.datetime(2014,9,1))]
return trainset, testset
3.1 使用SHAP包获取优胜特征¶
SHAP testues represent the fair score of features depending on their contribution towards the total score in the set of features.
SHAP also can visualize how the score changes when the feature testue is low/high on each data.
In [63]:
shap.initjs()
from sklearn import tree
model = tree.DecisionTreeRegressor()
train, test = split_data_underline(feature.dropna())
features = [x for x in train.columns if x not in date_indexs]
model.fit(train[features].drop(labels+['date'], axis=1), train['total_purchase_amt'])
explainer = shap.TreeExplainer(model)
shap_testues = explainer.shap_values(test[features].drop(labels+['date'], axis=1))
shap.summary_plot(shap_testues, test[features].drop(labels+['date'], axis=1), plot_type='bar')
shap.summary_plot(shap_testues, test[features].drop(labels+['date'], axis=1))
tree_important_purchase = pd.DataFrame(np.mean(np.abs(shap_testues), axis=0),[x for x in features if x not in labels + date_indexs + ['date']]).reset_index()
In [64]:
tree_important_purchase = tree_important_purchase.sort_values(0, ascending=False).reset_index(drop=True)
tree_important_purchase = list(tree_important_purchase[:20]['index'])
In [65]:
tree_important_purchase
Out[65]:
In [66]:
# 输出选择的特征
def draw_cloud(feature_index: List[str])->None:
plt.figure(figsize=(20,10))
plt.subplot(1,2,1)
ciyun = WordCloud(background_color='white', max_font_size=40)
ciyun.generate(text=''.join([x+' ' for x in feature_index if x != 'total_purchase_amt']))
plt.imshow(ciyun, interpolation='bilinear')
plt.axis("off")
In [67]:
draw_cloud(tree_important_purchase)
3.2 使用 Permutation Importance 包获取优胜特征
SHAP testues represent the fair score of features depending on their contribution towards the total score in the set of features.
SHAP also can visualize how the score changes when the feature testue is low/high on each data.
In [68]:
model = LinearRegression()
train, test = split_data_underline(feature.dropna())
model.fit(train[features].drop(labels+['date'], axis=1), train['total_purchase_amt'])
perm = PermutationImportance(model, random_state=42).fit(test[features].drop(labels+['date'], axis=1), test['total_purchase_amt'])
liner_important_purchase = pd.DataFrame(np.abs(perm.feature_importances_), [x for x in features if x not in labels + date_indexs + ['date']]).reset_index()
eli5.show_weights(perm, feature_names=list(str(x) for x in features if x not in labels + ['date']))
Out[68]:
In [69]:
liner_important_purchase = liner_important_purchase.sort_values(0, ascending=False).reset_index(drop=True)
liner_important_purchase = list(liner_important_purchase[:20]['index'])
In [70]:
draw_cloud(liner_important_purchase)
3.3 量特征集合取交集选出最终优胜特征¶
In [71]:
winer_features_purchase = list(set(tree_important_purchase)\
& set(liner_important_purchase))
In [72]:
draw_cloud(winer_features_purchase)
数据准备&辅助函数整合
In [73]:
import pandas as pd
import sklearn as skr
import numpy as np
import datetime
import matplotlib.pyplot as plt
import seaborn as sns
from dateutil.relativedelta import relativedelta
from typing import *
import random
from sklearn.metrics import mean_squared_error
from sklearn.metrics import mean_absolute_error
from sklearn.linear_model import LinearRegression
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.neural_network import MLPRegressor
import xgboost as xgb
import warnings
warnings.filterwarnings('ignore')
np.random.seed(1024)
labels = ['total_purchase_amt', 'total_redeem_amt']
In [74]:
# 分割数据集
def split_data_underline(data: pd.DataFrame)->pd.DataFrame:
trainset = data[(datetime.datetime(2014,4,1) <= data['date']) & (data['date'] < datetime.datetime(2014,8,1))]
testset = data[(datetime.datetime(2014,8,1) <= data['date']) & (data['date'] < datetime.datetime(2014,9,1))]
return trainset, testset
def split_data_online(data: pd.DataFrame)->pd.DataFrame:
trainset = data[(datetime.datetime(2014,4,1) <= data['date']) & (data['date'] < datetime.datetime(2014,9,1))]
testset = data[(datetime.datetime(2014,9,1) <= data['date']) & (data['date'] < datetime.datetime(2014,10,1))]
return trainset, testset
In [75]:
# 定义评价函数
def AE(y: Iterable, yhat: Iterable)->Iterable:
return np.abs(y - yhat) / np.abs(y)
def total_AE(purchasehat: Iterable, redeemhat: Iterable, purchase: Iterable, redeem: Iterable, h: int = 0.3)->Iterable:
return sum(map(lambda x : np.exp(-x/h)*10, AE(purchase, purchasehat))) * 0.45 + sum(map(lambda x : np.exp(-x/h)*10, AE(redeem, redeemhat))) * 0.55
In [76]:
# 在不同的时间段对模型进行验证
def week_evalution_single(data: pd.DataFrame, model: object, types: str)->pd.DataFrame:
results = []
a_month = relativedelta(months=1)
for i in [datetime.datetime(2014, 8, 1), datetime.datetime(2014, 7, 25), datetime.datetime(2014, 7, 18), datetime.datetime(2014, 7, 11),
datetime.datetime(2014, 7, 4), datetime.datetime(2014, 6, 27), datetime.datetime(2014, 6,20)]:
trainset = data[(i - 4 * a_month <= data['date']) & (data['date'] < i)]
testset = data[(i <= data['date']) & (data['date'] < i + a_month)]
if len(testset) == 0 or len(trainset) == 0:
i = datetime.datetime(2014, 4, 20)
trainset = data[(i - 4 * a_month <= data['date']) & (data['date'] < i)]
testset = data[(i <= data['date']) & (data['date'] < datetime.datetime(2014, 9, 1))]
feature = [x for x in trainset.columns if x not in ['total_purchase_amt','total_redeem_amt','date']]
model.fit(X=trainset[feature], y=trainset['total_' + types + '_amt'])
result_lr = model.predict(testset[feature])
h = 0.3
results.append(sum(AE(testset['total_' + types + '_amt'], result_lr).apply(lambda x : np.exp(-x/h))*10))
return pd.DataFrame(results)
In [77]:
# 输出评级表格
def draw_eva_table(df: pd.DataFrame)->pd.DataFrame:
rest = df.copy()
rest['interval'] = [datetime.date(2014, 8, 1), datetime.date(2014, 7, 25), datetime.date(2014, 7, 18), datetime.date(2014, 7, 11),
datetime.date(2014, 7, 4), datetime.date(2014, 6, 27), datetime.date(2014, 6,20)]
return rest
In [78]:
# 对生成结果进行可视化
def visual(result_purchase_lr: Iterable, result_redeem_lr: Iterable, testset: pd.DataFrame)->None:
fig = plt.figure(figsize=(10,4))
plt.plot(testset['date'], result_purchase_lr, label='predicted_purchase')
plt.plot(testset['date'], testset['total_purchase_amt'], label='real_redeem')
plt.legend(loc='best')
plt.title("The distribution of real and predict purchase")
plt.xlabel("Time")
plt.ylabel("Amount")
plt.show()
fig = plt.figure(figsize=(10,4))
sns.barplot(testset['date'].dt.day ,result_purchase_lr - testset['total_purchase_amt'])
fig = plt.figure(figsize=(10,4))
plt.plot(testset['date'], result_redeem_lr, label='predicted_redeem')
plt.plot(testset['date'], testset['total_redeem_amt'], label='real_redeem')
plt.legend(loc='best')
plt.title("The distribution of real and predict redeem")
plt.xlabel("Time")
plt.ylabel("Amount")
plt.show()
fig = plt.figure(figsize=(10,4))
sns.barplot(testset['date'].dt.day ,result_redeem_lr - testset['total_redeem_amt'])
In [79]:
# 定义提取线下最好效果特征的函数
def feature_extract(data: pd.DataFrame, model: object, types: str)->Tuple[List[str], List[float]]:
features = [x for x in data.columns if x not in labels + ['date']]
random.shuffle(features)
results = []
score = -1
for i in features:
score_update = np.mean(week_evalution_single(data[results + [i] + labels + ['date']], model, types))
if score_update > score:
score = score_update
results.append(i)
return results, score
def robust_feature_extract(data: pd.DataFrame, model: object, types: str):
results = []
score = -1
for i in range(10):
results_update, score_update = feature_extract(data, model, types)
if score_update > score:
score = score_update
results = results_update
print(results_update, score_update)
return results
In [80]:
# 定义AIC,BIC评价指标
def AIC(L: Iterable, delta: float, n_features: int):
return L * np.log10(delta) + 2 * (n_features + 1)
def AIC(L: Iterable, delta: float, n_features: int):
return L * np.log10(delta) + (n_features + 1) * np.log10(L)
In [81]:
# 使用AIC指标融合模型
def feature_extract_AIC(data: pd.DataFrame, model: object, types: str)->Tuple[List[str], float]:
features = [x for x in data.columns if x not in labels + ['date']]
random.shuffle(features)
results = []
test_score = 1e9
train_score = 0
for i in features:
test_score_update = np.mean(week_evalution_single(data[results + [i] + labels + ['date']], model, types)[0])
if test_score_update < test_score:
test_score = test_score_update
results.append(i)
trainset, testset = split_data_underline(data)
feature = results
model.fit(X=trainset[feature], y=trainset['total_' + types + '_amt'])
train_result_lr = model.predict(trainset[feature])
delta = mean_squared_error(train_result_lr, trainset['total_' + types + '_amt'])
#delta = np.sum(AE(trainset['total_' + types + '_amt'], train_result_lr).apply(lambda x : np.exp(-x/0.1))*10)
return results, AIC(len(trainset), delta, len(feature))
def multi_model(data: pd.DataFrame, model: object, types: str)->Tuple[List[List[str]], float]:
features = []
weights = []
for i in range(100):
results_update, score_update = feature_extract_AIC(data, model, types)
features.append(results_update)
weights.append(score_update)
avg = np.mean(weights)
weights = [x - avg for x in weights]
weights = [np.power((-1 * x / 2), 10) for x in weights]
summ = np.sum(weights)
weights = [x / summ for x in weights]
return features, weights
In [82]:
# 生成线上结果
def generate_online_result(df: pd.DataFrame, feature: Iterable, model = LinearRegression(), target:str = 'total_purchase_amt')->Iterable:
trainset, testset = split_data_online(df)
model.fit(X=trainset[feature], y=trainset[target])
result_purchase_lr = model.predict(testset[feature])
return result_purchase_lr
In [83]:
def generate_under_result(df: pd.DataFrame, feature: Iterable, model = LinearRegression(), target:str = 'total_purchase_amt')->Iterable:
trainset, testset = split_data_underline(df)
model.fit(X=trainset[feature], y=trainset[target])
result_purchase_lr = model.predict(testset[feature])
return result_purchase_lr
In [84]:
# 生成线上提交的格式
def normalize_upload_file(result_purchase_lr: Iterable, result_redeem_lr: Iterable, testset: pd.DataFrame)->pd.DataFrame:
testset['total_purchase_amt'] = result_purchase_lr
testset['total_redeem_amt'] = result_redeem_lr
online_upload = testset[['date','total_purchase_amt','total_redeem_amt']]
online_upload['date'] = online_upload['date'].astype(str)
online_upload['date'] = online_upload['date'].str.replace('-','')
return online_upload
In [85]:
# 线上结果可视化
def draw_result(result_purchase_lr: Iterable, result_redeem_lr: Iterable, testset: pd.DataFrame):
fig = plt.figure(figsize=(12,6))
plt.plot(testset['date'], result_purchase_lr, label='online_purchase', linewidth = 0.7, linestyle = '--', marker = 'o')
plt.plot(testset['date'], result_redeem_lr, label='online_redeem', linewidth = 1.2, marker = '*')
plt.legend(loc='best')
plt.title("The predict values")
plt.xlabel("Time")
plt.ylabel("Amount")
In [86]:
# 重载DataFrame加法
def add_two_df(df1, df2, features = None, left_a = 0.45, right_a = 0.55):
data = df1.copy()
if not features:
features = [x for x in data.columns if x != 'interval']
for i in features:
data[i] = (data[i] * left_a + df2[i] * right_a)
return data
In [87]:
# 重载DataFrame乘法
def scale_df(df1, features = None, eta = 1):
data = df1.copy()
if not features:
features = [x for x in data.columns if x != 'interval']
for i in features:
data[i] *= eta
return data
1. 仅用 is 特征建模¶
In [88]:
winer_features_purchase
Out[88]:
In [89]:
data = pd.read_csv('Feature/purchase_feature_droped_0614.csv')
data['date'] = pd.to_datetime(data['date'])
In [90]:
trainset, testset = split_data_underline(data)
result_purchase_lr = generate_under_result(data, [x for x in data.columns if x not in ['total_purchase_amt','total_redeem_amt','date']], target='total_purchase_amt')
result_redeem_lr = generate_under_result(data, [x for x in data.columns if x not in ['total_purchase_amt','total_redeem_amt','date']], target='total_redeem_amt')
1.1 八月预测¶
In [91]:
total_AE(result_purchase_lr, result_redeem_lr, testset['total_purchase_amt'], testset['total_redeem_amt'])
Out[91]:
In [92]:
draw_eva_table(week_evalution_single(data, model=LinearRegression(), types = 'purchase'))
Out[92]:
1.2 滑动窗口测试¶
In [93]:
draw_eva_table(week_evalution_single(data, LinearRegression(), 'redeem'))
Out[93]:
1.3 对比分析¶
In [94]:
visual(result_purchase_lr, result_redeem_lr, testset)
In [95]:
result_purchase_lr = generate_online_result(data, [x for x in trainset.columns if x not in ['total_purchase_amt','total_redeem_amt','date']], LinearRegression(),'total_purchase_amt')
result_redeem_lr = generate_online_result(data, [x for x in trainset.columns if x not in ['total_purchase_amt','total_redeem_amt','date']], LinearRegression(),'total_redeem_amt')
1.4 九月预测¶
In [96]:
trainset, testset = split_data_online(data)
draw_result(result_purchase_lr, result_redeem_lr, testset)
In [97]:
normalize_upload_file(result_purchase_lr, result_redeem_lr, testset).to_csv('20190612_only_is.csv',index=False,header=None)
2. 多模型对比分析¶
In [98]:
def multi_model_eva(data, types:str = 'purchase'):
results = pd.DataFrame()
for model in [LinearRegression(), DecisionTreeRegressor(), RandomForestRegressor(), GradientBoostingRegressor(), MLPRegressor(solver='lbfgs'), xgb.XGBRegressor(objective='reg:squarederror')]:
if results.empty:
results = draw_eva_table(week_evalution_single(data, model, types)).rename(columns={0: repr(model).split('(')[0]})
else:
results = pd.merge(results, \
draw_eva_table(week_evalution_single(data, model, types)).rename(columns={0: repr(model).split('(')[0]}), on='interval')
results = results[['interval'] + [x for x in results.columns if x != 'interval']]
return results
In [99]:
multi_result = add_two_df(multi_model_eva(data, 'purchase'), multi_model_eva(data, 'redeem')).set_index('interval')
multi_result
Out[99]:
In [100]:
multi_result.plot(figsize = (12,5),kind = 'line', marker = 'o', linewidth = 0.7, linestyle = '--')
plt.title('多模型对比分析')
plt.xlabel('时间间隔')
plt.ylabel('分数')
Out[100]:
神经网络模型¶
In [101]:
import math
import numpy
import pandas
from keras.layers import LSTM, RNN, GRU, SimpleRNN
from keras.layers import Dense, Dropout
from keras.callbacks import EarlyStopping
import matplotlib.pyplot as plt
from keras.models import Sequential
from sklearn.preprocessing import MinMaxScaler
import os
numpy.random.seed(2019)
class RNNModel(object):
def __init__(self, look_back=1, epochs_purchase=20, epochs_redeem=40, batch_size=1, verbose=2, patience=10, store_result=False):
self.look_back = look_back
self.epochs_purchase = epochs_purchase
self.epochs_redeem = epochs_redeem
self.batch_size = batch_size
self.verbose = verbose
self.store_result = store_result
self.patience = patience
self.purchase = pandas.read_csv('Dataset/date_label.csv', usecols=[1], engine='python')
self.redeem = pandas.read_csv('Dataset/date_label.csv', usecols=[2], engine='python')
def access_data(self, data_frame):
# load the data set
data_set = data_frame.values
data_set = data_set.astype('float32')
# LSTMs are sensitive to the scale of the input data, specifically when the sigmoid (default) or tanh activation functions are used. It can be a good practice to rescale the data to the range of 0-to-1, also called normalizing.
scaler = MinMaxScaler(feature_range=(0, 1))
data_set = scaler.fit_transform(data_set)
# reshape into X=t and Y=t+1
train_x, train_y, test = self.create_data_set(data_set)
# reshape input to be [samples, time steps, features]
train_x = numpy.reshape(train_x, (train_x.shape[0], 1, train_x.shape[1]))
return train_x, train_y, test, scaler
# convert an array of values into a data set matrix
def create_data_set(self, data_set):
data_x, data_y = [], []
for i in range(len(data_set)-self.look_back - 30):
a = data_set[i:(i + self.look_back), 0]
data_x.append(a)
data_y.append(list(data_set[i + self.look_back: i + self.look_back + 30, 0]))
# print(numpy.array(data_y).shape)
return numpy.array(data_x), numpy.array(data_y), data_set[-self.look_back:, 0].reshape(1, 1, self.look_back)
def rnn_model(self, train_x, train_y, epochs):
model = Sequential()
model.add(LSTM(64, input_shape=(1, self.look_back), return_sequences=True))
model.add(LSTM(32, return_sequences=False))
model.add(Dense(32))
model.add(Dense(30))
#model.add(Dense(30))
model.compile(loss='mean_squared_error', optimizer='adam')
model.summary()
early_stopping = EarlyStopping('loss', patience=self.patience)
history = model.fit(train_x, train_y, epochs=epochs, batch_size=self.batch_size, verbose=self.verbose, callbacks=[early_stopping])
return model
def predict(self, model, data):
prediction = model.predict(data)
return prediction
def plot_show(self, predict):
predict = predict[['purchase', 'redeem']]
plt.figure(figsize = (12,8))
predict.plot(linewidth = 0.8, marker = 'o', linestyle = '--', markersize=6)
plt.title('2014年9月预测对比分析：LSTM模型可视化')
plt.xlabel('时间')
plt.ylabel('金额')
new_xticks = ['9.' + str(x) for x in range(1,31)]
plt.xticks(np.arange(30),new_xticks, rotation = 45)
plt.legend(bbox_to_anchor=(1.05, 0), loc=3, borderaxespad=0)
def run(self):
purchase_train_x, purchase_train_y, purchase_test, purchase_scaler = self.access_data(self.purchase)
redeem_train_x, redeem_train_y, redeem_test, redeem_scaler = self.access_data(self.redeem)
purchase_model = self.rnn_model(purchase_train_x, purchase_train_y, self.epochs_purchase)
redeem_model = self.rnn_model(redeem_train_x, redeem_train_y, self.epochs_redeem)
purchase_predict = self.predict(purchase_model, purchase_test)
redeem_predict = self.predict(redeem_model, redeem_test)
test_user = pandas.DataFrame({'report_date': [20140900 + i for i in range(1, 31)]})
purchase = purchase_scaler.inverse_transform(purchase_predict).reshape(30, 1)
redeem = redeem_scaler.inverse_transform(redeem_predict).reshape(30, 1)
test_user['purchase'] = purchase
test_user['redeem'] = redeem
print(test_user)
"""Store submit file"""
if self.store_result is True:
test_user.to_csv('Dataset/submit_lstm.csv', encoding='utf-8', index=None, header=None)
"""plot result picture"""
self.plot_show(test_user)
if __name__ == '__main__':
initiation = RNNModel(look_back=40, epochs_purchase=150, epochs_redeem=230, batch_size=16, verbose=1, patience=50, store_result=False)
initiation.run()
劣汰后特征对比 ¶
In [102]:
data_purchase = pd.read_csv('Feature/purchase_feature_droped_0614.csv')
data_purchase['date'] = pd.to_datetime(data_purchase['date'])
In [103]:
data_redeem = pd.read_csv('Feature/redeem_feature_droped_0614.csv')
data_redeem['date'] = pd.to_datetime(data_redeem['date'])
In [104]:
trainset_purchase, testset_purchase = split_data_underline(data_purchase)
result_purchase_lr = generate_under_result(data_purchase, [x for x in data_purchase.columns
if x not in ['total_purchase_amt','total_redeem_amt','date']],
target='total_purchase_amt')
In [105]:
trainset_redeem, testset_redeem = split_data_underline(data_redeem)
result_redeem_lr = generate_under_result(data_redeem, [x for x in data_redeem.columns
if x not in ['total_purchase_amt','total_redeem_amt','date']],
target='total_redeem_amt')
In [106]:
total_AE(result_purchase_lr, result_redeem_lr, testset_purchase['total_purchase_amt'], testset_redeem['total_redeem_amt'])
Out[106]:
In [107]:
add_two_df(multi_model_eva(data_purchase, 'purchase'), multi_model_eva(data_redeem, 'redeem'))
Out[107]:
In [108]:
trainset, testset = split_data_underline(data)
visual(result_purchase_lr, result_redeem_lr, testset)
In [109]:
result_purchase_lr = generate_online_result(data_purchase, [x for x in data_purchase.columns if x not in ['total_purchase_amt','total_redeem_amt','date']], LinearRegression(),'total_purchase_amt')
result_redeem_lr = generate_online_result(data_redeem, [x for x in data_redeem.columns if x not in ['total_purchase_amt','total_redeem_amt','date']], LinearRegression(),'total_redeem_amt')
3.1 线性回归¶
In [110]:
trainset, testset = split_data_online(data)
draw_result(result_purchase_lr, result_redeem_lr, testset)
purchase feature
'dis_to_nowork', 'dis_to_work', 'dis_from_work', 'purchase_weekdayrate', 'redeem_dayrate', 'weekday_onehot_5', 'weekday_onehot_6', 'dis_from_nowork', 'is_holiday', 'weekday_onehot_1', 'weekday_onehot_2', 'weekday_onehot_0', 'dis_from_middleofweek', 'dis_from_holiendday', 'weekday_onehot_3', 'is_lastday_of_holiday', 'is_firstday_of_holiday', 'weekday_onehot_4', 'is_worked_yestday', 'is_second_week', 'is_third_week', 'dis_from_startofmonth', 'dis_from_holiday', 'dis_to_nowork%%%%dis_from_purchase_peak', 'total_purchase_amt', 'total_redeem_amt', 'date'
Redeem feature
'is_work', 'dis_from_redeem_valley', 'purchase_weekdayrate', 'redeem_dayrate', 'weekday_onehot_5', 'is_gonna_work_tomorrow', 'is_holiday', 'dis_from_nowork', 'weekday_onehot_0', 'weekday_onehot_1', 'is_firstday_of_holiday', 'weekday_onehot_2', 'is_lastday_of_holiday', 'dis_from_holiday', 'is_work_on_sunday', 'is_firstday_of_work', 'is_secday_of_month', 'dis_from_holiendday', 'dis_from_redeem_valley%%%%dis_from_redeem_peak', 'total_purchase_amt', 'total_redeem_amt', 'date'
In [111]:
normalize_upload_file(result_purchase_lr, result_redeem_lr, testset).to_csv('Dataset/20190614_droped.csv',index=False,header=None)
3.2 MLP¶
In [112]:
result_purchase_lr = generate_online_result(data_purchase, [x for x in data_purchase.columns
if x not in ['total_purchase_amt','total_redeem_amt','date']],
MLPRegressor(solver='lbfgs'),'total_purchase_amt')
result_redeem_lr = generate_online_result(data_redeem, [x for x in data_redeem.columns
if x not in ['total_purchase_amt','total_redeem_amt','date']],
MLPRegressor(solver='lbfgs'),'total_redeem_amt')
trainset, testset = split_data_online(data)
draw_result(result_purchase_lr, result_redeem_lr, testset)
In [113]:
normalize_upload_file(result_purchase_lr, result_redeem_lr, testset).to_csv('Dataset/20190614_droped_MLP.csv',index=False,header=None)
3.3 Xgboost¶
In [114]:
result_purchase_lr = generate_online_result(data_purchase, [x for x in data_purchase.columns
if x not in ['total_purchase_amt','total_redeem_amt','date']],
xgb.XGBRegressor(objective='reg:squarederror'),'total_purchase_amt')
result_redeem_lr = generate_online_result(data_redeem, [x for x in data_redeem.columns
if x not in ['total_purchase_amt','total_redeem_amt','date']],
xgb.XGBRegressor(objective='reg:squarederror'),'total_redeem_amt')
trainset, testset = split_data_online(data)
draw_result(result_purchase_lr, result_redeem_lr, testset)
In [115]:
normalize_upload_file(result_purchase_lr, result_redeem_lr, testset).to_csv('Dataset/20190615_droped_XGB.csv',index=False,header=None)