1. hotel 데이터셋 살펴보기

In [ ]:

import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

In [ ]:

hotel_df = pd.read_csv('/content/drive/MyDrive/KDT/4. 머신러닝과 딥러닝/hotel.csv')

In [ ]:

hotel_df

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	...	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	reservation_status_date	name	email	phone-number	credit_card
0	Resort Hotel	0	342	2015	July	27	1	0	0	2	...	0	Transient	0.00	0	0	2015-07-01	Ernest Barnes	Ernest.Barnes31@outlook.com	669-792-1661	************4322
1	Resort Hotel	0	737	2015	July	27	1	0	0	2	...	0	Transient	0.00	0	0	2015-07-01	Andrea Baker	Andrea_Baker94@aol.com	858-637-6955	************9157
2	Resort Hotel	0	7	2015	July	27	1	0	1	1	...	0	Transient	75.00	0	0	2015-07-02	Rebecca Parker	Rebecca_Parker@comcast.net	652-885-2745	************3734
3	Resort Hotel	0	13	2015	July	27	1	0	1	1	...	0	Transient	75.00	0	0	2015-07-02	Laura Murray	Laura_M@gmail.com	364-656-8427	************5677
4	Resort Hotel	0	14	2015	July	27	1	0	2	2	...	0	Transient	98.00	0	1	2015-07-03	Linda Hines	LHines@verizon.com	713-226-5883	************5498
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
119385	City Hotel	0	23	2017	August	35	30	2	5	2	...	0	Transient	96.14	0	0	2017-09-06	Claudia Johnson	Claudia.J@yahoo.com	403-092-5582	************8647
119386	City Hotel	0	102	2017	August	35	31	2	5	3	...	0	Transient	225.43	0	2	2017-09-07	Wesley Aguilar	WAguilar@xfinity.com	238-763-0612	************4333
119387	City Hotel	0	34	2017	August	35	31	2	5	2	...	0	Transient	157.71	0	4	2017-09-07	Mary Morales	Mary_Morales@hotmail.com	395-518-4100	************1821
119388	City Hotel	0	109	2017	August	35	31	2	5	2	...	0	Transient	104.40	0	0	2017-09-07	Caroline Conley MD	MD_Caroline@comcast.net	531-528-1017	************7860
119389	City Hotel	0	205	2017	August	35	29	2	7	2	...	0	Transient	151.20	0	2	2017-09-07	Ariana Michael	Ariana_M@xfinity.com	422-804-6403	************4482

119390 rows × 32 columns

In [ ]:

pd.set_option('display.max_columns', 100)

In [ ]:

hotel_df.head()

Out[ ]:

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	meal	country	distribution_channel	reserved_room_type	assigned_room_type	booking_changes	deposit_type	customer_type	adr	total_of_special_requests	reservation_status_date	name	email	phone-number	credit_card
0	Resort Hotel	342	2015	July	27	1	0	2	BB	PRT	Direct	C	C	3	No Deposit	Transient	0.0	0	2015-07-01	Ernest Barnes	Ernest.Barnes31@outlook.com	669-792-1661	************4322
1	Resort Hotel	737	2015	July	27	1	0	2	BB	PRT	Direct	C	C	4	No Deposit	Transient	0.0	0	2015-07-01	Andrea Baker	Andrea_Baker94@aol.com	858-637-6955	************9157
2	Resort Hotel	7	2015	July	27	1	1	1	BB	GBR	Direct	A	C	0	No Deposit	Transient	75.0	0	2015-07-02	Rebecca Parker	Rebecca_Parker@comcast.net	652-885-2745	************3734
3	Resort Hotel	13	2015	July	27	1	1	1	BB	GBR	Corporate	A	A	0	No Deposit	Transient	75.0	0	2015-07-02	Laura Murray	Laura_M@gmail.com	364-656-8427	************5677
4	Resort Hotel	14	2015	July	27	1	2	2	BB	GBR	TA/TO	A	A	0	No Deposit	Transient	98.0	1	2015-07-03	Linda Hines	LHines@verizon.com	713-226-5883	************5498

In [ ]:

hotel_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 119390 entries, 0 to 119389
Data columns (total 32 columns):
 #   Column                          Non-Null Count   Dtype  
---  ------                          --------------   -----  
 0   hotel                           119390 non-null  object 
 1   is_canceled                     119390 non-null  int64  
 2   lead_time                       119390 non-null  int64  
 3   arrival_date_year               119390 non-null  int64  
 4   arrival_date_month              119390 non-null  object 
 5   arrival_date_week_number        119390 non-null  int64  
 6   arrival_date_day_of_month       119390 non-null  int64  
 7   stays_in_weekend_nights         119390 non-null  int64  
 8   stays_in_week_nights            119390 non-null  int64  
 9   adults                          119390 non-null  int64  
 10  children                        119386 non-null  float64
 11  babies                          119390 non-null  int64  
 12  meal                            119390 non-null  object 
 13  country                         118902 non-null  object 
 14  distribution_channel            119390 non-null  object 
 15  is_repeated_guest               119390 non-null  int64  
 16  previous_cancellations          119390 non-null  int64  
 17  previous_bookings_not_canceled  119390 non-null  int64  
 18  reserved_room_type              119390 non-null  object 
 19  assigned_room_type              119390 non-null  object 
 20  booking_changes                 119390 non-null  int64  
 21  deposit_type                    119390 non-null  object 
 22  days_in_waiting_list            119390 non-null  int64  
 23  customer_type                   119390 non-null  object 
 24  adr                             119390 non-null  float64
 25  required_car_parking_spaces     119390 non-null  int64  
 26  total_of_special_requests       119390 non-null  int64  
 27  reservation_status_date         119390 non-null  object 
 28  name                            119390 non-null  object 
 29  email                           119390 non-null  object 
 30  phone-number                    119390 non-null  object 
 31  credit_card                     119390 non-null  object 
dtypes: float64(2), int64(16), object(14)
memory usage: 29.1+ MB

hotel: 호텔 종류
is_canceled: 취소 여부
lead_time: 예약 시점으로부터 체크인 될 때까지의 기간(얼마나 미리 예약했는지)
arrival_date_year: 예약 연도
arrival_date_month: 예약 월
arrival_date_week_number: 예약 주
arrival_date_day_of_month: 예약 일
stays_in_weekend_nights: 주말을 끼고 얼마나 묵었는지
stays_in_week_nights: 평일을 끼고 얼마나 묵었는지
adults: 성인 인원수
children: 어린이 인원수
babies: 아기 인원수
meal: 식사 형태
country: 지역
distribution_channel: 어떤 방식으로 예약했는지
is_repeated_guest: 예약한적이 있는 고객인지
previous_cancellations: 몇번 예약을 취소했었는지
previous_bookings_not_canceled: 예약을 취소하지 않고 정상 숙박한 횟수
reserved_room_type: 희망한 룸타입
assigned_room_type: 실제 배정된 룸타입
booking_changes: 예약 후 서비스가 몇번 변경되었는지
deposit_type: 요금 납부 방식
days_in_waiting_list: 예약을 위해 기다린 날짜
customer_type: 고객 타입
adr: 특정일에 높아지거나 낮아지는 가격
required_car_parking_spaces: 주차공간을 요구했는지
total_of_special_requests: 특별한 별도의 요청사항이 있는지
reservation_status_date: 예약한 날짜
name: 이름
email: 이메일
phone-number: 전화번호
credit_card: 카드번호

In [ ]:

hotel_df.drop(['credit_card', 'email', 'name', 'phone-number', 'reservation_status_date'], axis=1, inplace=True)

In [ ]:

hotel_df.head()

Out[ ]:

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	meal	country	distribution_channel	reserved_room_type	assigned_room_type	booking_changes	deposit_type	customer_type	adr	total_of_special_requests
0	Resort Hotel	342	2015	July	27	1	0	2	BB	PRT	Direct	C	C	3	No Deposit	Transient	0.0	0
1	Resort Hotel	737	2015	July	27	1	0	2	BB	PRT	Direct	C	C	4	No Deposit	Transient	0.0	0
2	Resort Hotel	7	2015	July	27	1	1	1	BB	GBR	Direct	A	C	0	No Deposit	Transient	75.0	0
3	Resort Hotel	13	2015	July	27	1	1	1	BB	GBR	Corporate	A	A	0	No Deposit	Transient	75.0	0
4	Resort Hotel	14	2015	July	27	1	2	2	BB	GBR	TA/TO	A	A	0	No Deposit	Transient	98.0	1

In [ ]:

hotel_df.describe()

Out[ ]:

	is_canceled	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	booking_changes	days_in_waiting_list	adr	required_car_parking_spaces	total_of_special_requests
count	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119386.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000	119390.000000
mean	0.370416	104.011416	2016.156554	27.165173	15.798241	0.927599	2.500302	1.856403	0.103890	0.007949	0.031912	0.087118	0.137097	0.221124	2.321149	101.831122	0.062518	0.571363
std	0.482918	106.863097	0.707476	13.605138	8.780829	0.998613	1.908286	0.579261	0.398561	0.097436	0.175767	0.844336	1.497437	0.652306	17.594721	50.535790	0.245291	0.792798
min	0.000000	0.000000	2015.000000	1.000000	1.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	-6.380000	0.000000	0.000000
25%	0.000000	18.000000	2016.000000	16.000000	8.000000	0.000000	1.000000	2.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	69.290000	0.000000	0.000000
50%	0.000000	69.000000	2016.000000	28.000000	16.000000	1.000000	2.000000	2.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	94.575000	0.000000	0.000000
75%	1.000000	160.000000	2017.000000	38.000000	23.000000	2.000000	3.000000	2.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	126.000000	0.000000	1.000000
max	1.000000	737.000000	2017.000000	53.000000	31.000000	19.000000	50.000000	55.000000	10.000000	10.000000	1.000000	26.000000	72.000000	21.000000	391.000000	5400.000000	8.000000	5.000000

In [ ]:

sns.displot(hotel_df['lead_time'])

Out[ ]:

<seaborn.axisgrid.FacetGrid at 0x7f5e470750f0>

In [ ]:

sns.boxplot(y=hotel_df['lead_time'])

Out[ ]:

<Axes: ylabel='lead_time'>

In [ ]:

sns.barplot(x=hotel_df['distribution_channel'], y=hotel_df['is_canceled'])

Out[ ]:

<Axes: xlabel='distribution_channel', ylabel='is_canceled'>

In [ ]:

hotel_df['distribution_channel'].value_counts()

Out[ ]:

TA/TO        97870
Direct       14645
Corporate     6677
GDS            193
Undefined        5
Name: distribution_channel, dtype: int64

In [ ]:

sns.barplot(x=hotel_df['hotel'], y=hotel_df['is_canceled'])

Out[ ]:

<Axes: xlabel='hotel', ylabel='is_canceled'>

In [ ]:

sns.barplot(x=hotel_df['arrival_date_year'], y=hotel_df['is_canceled'])

Out[ ]:

<Axes: xlabel='arrival_date_year', ylabel='is_canceled'>

In [ ]:

plt.figure(figsize=(15, 5))
sns.barplot(x=hotel_df['arrival_date_month'], y=hotel_df['is_canceled'])

Out[ ]:

<Axes: xlabel='arrival_date_month', ylabel='is_canceled'>

In [ ]:

import calendar

In [ ]:

print(calendar.month_name[1])
print(calendar.month_name[2])
print(calendar.month_name[3])

January
February
March

In [ ]:

months = []

for i in range(1, 13):
    months.append(calendar.month_name[i])

In [ ]:

months

Out[ ]:

['January',
 'February',
 'March',
 'April',
 'May',
 'June',
 'July',
 'August',
 'September',
 'October',
 'November',
 'December']

In [ ]:

plt.figure(figsize=(15, 5))
sns.barplot(x=hotel_df['arrival_date_month'], y=hotel_df['is_canceled'], order=months)

Out[ ]:

<Axes: xlabel='arrival_date_month', ylabel='is_canceled'>

In [ ]:

sns.barplot(x=hotel_df['is_repeated_guest'], y=hotel_df['is_canceled'])

Out[ ]:

<Axes: xlabel='is_repeated_guest', ylabel='is_canceled'>

In [ ]:

sns.barplot(x=hotel_df['deposit_type'], y=hotel_df['is_canceled'])

Out[ ]:

<Axes: xlabel='deposit_type', ylabel='is_canceled'>

In [ ]:

hotel_df['deposit_type'].value_counts()

Out[ ]:

No Deposit    104641
Non Refund     14587
Refundable       162
Name: deposit_type, dtype: int64

In [ ]:

plt.figure(figsize=(15, 15))
sns.heatmap(hotel_df.corr(), cmap='coolwarm', vmax=1, vmin=-1, annot=True)

<ipython-input-25-e4a176d0acdd>:2: FutureWarning: The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.
  sns.heatmap(hotel_df.corr(), cmap='coolwarm', vmax=1, vmin=-1, annot=True)

Out[ ]:

<Axes: >

In [ ]:

hotel_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 119390 entries, 0 to 119389
Data columns (total 27 columns):
 #   Column                          Non-Null Count   Dtype  
---  ------                          --------------   -----  
 0   hotel                           119390 non-null  object 
 1   is_canceled                     119390 non-null  int64  
 2   lead_time                       119390 non-null  int64  
 3   arrival_date_year               119390 non-null  int64  
 4   arrival_date_month              119390 non-null  object 
 5   arrival_date_week_number        119390 non-null  int64  
 6   arrival_date_day_of_month       119390 non-null  int64  
 7   stays_in_weekend_nights         119390 non-null  int64  
 8   stays_in_week_nights            119390 non-null  int64  
 9   adults                          119390 non-null  int64  
 10  children                        119386 non-null  float64
 11  babies                          119390 non-null  int64  
 12  meal                            119390 non-null  object 
 13  country                         118902 non-null  object 
 14  distribution_channel            119390 non-null  object 
 15  is_repeated_guest               119390 non-null  int64  
 16  previous_cancellations          119390 non-null  int64  
 17  previous_bookings_not_canceled  119390 non-null  int64  
 18  reserved_room_type              119390 non-null  object 
 19  assigned_room_type              119390 non-null  object 
 20  booking_changes                 119390 non-null  int64  
 21  deposit_type                    119390 non-null  object 
 22  days_in_waiting_list            119390 non-null  int64  
 23  customer_type                   119390 non-null  object 
 24  adr                             119390 non-null  float64
 25  required_car_parking_spaces     119390 non-null  int64  
 26  total_of_special_requests       119390 non-null  int64  
dtypes: float64(2), int64(16), object(9)
memory usage: 24.6+ MB

In [ ]:

hotel_df = hotel_df.dropna()

In [ ]:

hotel_df

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests
0	Resort Hotel	0	342	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	3	No Deposit	0	Transient	0.00	0	0
1	Resort Hotel	0	737	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	4	No Deposit	0	Transient	0.00	0	0
2	Resort Hotel	0	7	2015	July	27	1	0	1	1	0.0	0	BB	GBR	Direct	0	0	0	A	C	0	No Deposit	0	Transient	75.00	0	0
3	Resort Hotel	0	13	2015	July	27	1	0	1	1	0.0	0	BB	GBR	Corporate	0	0	0	A	A	0	No Deposit	0	Transient	75.00	0	0
4	Resort Hotel	0	14	2015	July	27	1	0	2	2	0.0	0	BB	GBR	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	98.00	0	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
119385	City Hotel	0	23	2017	August	35	30	2	5	2	0.0	0	BB	BEL	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	96.14	0	0
119386	City Hotel	0	102	2017	August	35	31	2	5	3	0.0	0	BB	FRA	TA/TO	0	0	0	E	E	0	No Deposit	0	Transient	225.43	0	2
119387	City Hotel	0	34	2017	August	35	31	2	5	2	0.0	0	BB	DEU	TA/TO	0	0	0	D	D	0	No Deposit	0	Transient	157.71	0	4
119388	City Hotel	0	109	2017	August	35	31	2	5	2	0.0	0	BB	GBR	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	104.40	0	0
119389	City Hotel	0	205	2017	August	35	29	2	7	2	0.0	0	HB	DEU	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	151.20	0	2

118898 rows × 27 columns

In [ ]:

hotel_df[hotel_df['adults'] == 0]

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests
2224	Resort Hotel	0	1	2015	October	41	6	0	3	0	0.0	0	SC	PRT	Corporate	0	0	0	A	I	1	No Deposit	0	Transient-Party	0.00	0	0
2409	Resort Hotel	0	0	2015	October	42	12	0	0	0	0.0	0	SC	PRT	Corporate	0	0	0	A	I	0	No Deposit	0	Transient	0.00	0	0
3181	Resort Hotel	0	36	2015	November	47	20	1	2	0	0.0	0	SC	ESP	TA/TO	0	0	0	A	C	0	No Deposit	0	Transient-Party	0.00	0	0
3684	Resort Hotel	0	165	2015	December	53	30	1	4	0	0.0	0	SC	PRT	TA/TO	0	0	0	A	A	1	No Deposit	122	Transient-Party	0.00	0	0
3708	Resort Hotel	0	165	2015	December	53	30	2	4	0	0.0	0	SC	PRT	TA/TO	0	0	0	A	C	1	No Deposit	122	Transient-Party	0.00	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
117204	City Hotel	0	296	2017	July	30	27	1	3	0	2.0	0	BB	GBR	TA/TO	0	0	0	B	A	0	No Deposit	0	Transient	98.85	0	1
117274	City Hotel	0	276	2017	July	31	30	2	1	0	2.0	0	BB	DEU	TA/TO	0	0	0	B	B	1	No Deposit	0	Transient	93.64	0	2
117303	City Hotel	0	291	2017	July	30	29	2	2	0	2.0	0	BB	PRT	TA/TO	0	0	0	B	A	0	No Deposit	0	Transient	98.85	0	1
117453	City Hotel	0	159	2017	July	31	31	1	3	0	2.0	0	SC	FRA	TA/TO	0	0	0	A	A	1	No Deposit	0	Transient	121.88	0	1
118200	City Hotel	0	10	2017	August	32	12	2	2	0	3.0	0	BB	MAR	Direct	0	0	0	B	A	1	No Deposit	0	Transient-Party	6.00	0	1

393 rows × 27 columns

In [ ]:

hotel_df['people'] = hotel_df['adults'] + hotel_df['children'] + hotel_df['babies']

<ipython-input-30-7b8904ba1394>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df['people'] = hotel_df['adults'] + hotel_df['children'] + hotel_df['babies']

In [ ]:

hotel_df.head()

Out[ ]:

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	meal	country	distribution_channel	reserved_room_type	assigned_room_type	booking_changes	deposit_type	customer_type	adr	total_of_special_requests	people
0	Resort Hotel	342	2015	July	27	1	0	2	BB	PRT	Direct	C	C	3	No Deposit	Transient	0.0	0	2.0
1	Resort Hotel	737	2015	July	27	1	0	2	BB	PRT	Direct	C	C	4	No Deposit	Transient	0.0	0	2.0
2	Resort Hotel	7	2015	July	27	1	1	1	BB	GBR	Direct	A	C	0	No Deposit	Transient	75.0	0	1.0
3	Resort Hotel	13	2015	July	27	1	1	1	BB	GBR	Corporate	A	A	0	No Deposit	Transient	75.0	0	1.0
4	Resort Hotel	14	2015	July	27	1	2	2	BB	GBR	TA/TO	A	A	0	No Deposit	Transient	98.0	1	2.0

In [ ]:

hotel_df[hotel_df['people'] == 0]

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	people
2224	Resort Hotel	0	1	2015	October	41	6	0	3	0	0.0	0	SC	PRT	Corporate	0	0	0	A	I	1	No Deposit	0	Transient-Party	0.00	0	0	0.0
2409	Resort Hotel	0	0	2015	October	42	12	0	0	0	0.0	0	SC	PRT	Corporate	0	0	0	A	I	0	No Deposit	0	Transient	0.00	0	0	0.0
3181	Resort Hotel	0	36	2015	November	47	20	1	2	0	0.0	0	SC	ESP	TA/TO	0	0	0	A	C	0	No Deposit	0	Transient-Party	0.00	0	0	0.0
3684	Resort Hotel	0	165	2015	December	53	30	1	4	0	0.0	0	SC	PRT	TA/TO	0	0	0	A	A	1	No Deposit	122	Transient-Party	0.00	0	0	0.0
3708	Resort Hotel	0	165	2015	December	53	30	2	4	0	0.0	0	SC	PRT	TA/TO	0	0	0	A	C	1	No Deposit	122	Transient-Party	0.00	0	0	0.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
115029	City Hotel	0	107	2017	June	26	27	0	3	0	0.0	0	BB	CHE	TA/TO	0	0	0	A	A	1	No Deposit	0	Transient	100.80	0	0	0.0
115091	City Hotel	0	1	2017	June	26	30	0	1	0	0.0	0	SC	PRT	Direct	0	0	0	E	K	0	No Deposit	0	Transient	0.00	1	1	0.0
116251	City Hotel	0	44	2017	July	28	15	1	1	0	0.0	0	SC	SWE	TA/TO	0	0	0	A	K	2	No Deposit	0	Transient	73.80	0	0	0.0
116534	City Hotel	0	2	2017	July	28	15	2	5	0	0.0	0	SC	RUS	TA/TO	0	0	0	A	K	1	No Deposit	0	Transient-Party	22.86	0	1	0.0
117087	City Hotel	0	170	2017	July	30	27	0	2	0	0.0	0	BB	BRA	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	0.00	0	0	0.0

170 rows × 28 columns

In [ ]:

hotel_df = hotel_df[hotel_df['people'] != 0]

In [ ]:

hotel_df

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	people
0	Resort Hotel	0	342	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	3	No Deposit	0	Transient	0.00	0	0	2.0
1	Resort Hotel	0	737	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	4	No Deposit	0	Transient	0.00	0	0	2.0
2	Resort Hotel	0	7	2015	July	27	1	0	1	1	0.0	0	BB	GBR	Direct	0	0	0	A	C	0	No Deposit	0	Transient	75.00	0	0	1.0
3	Resort Hotel	0	13	2015	July	27	1	0	1	1	0.0	0	BB	GBR	Corporate	0	0	0	A	A	0	No Deposit	0	Transient	75.00	0	0	1.0
4	Resort Hotel	0	14	2015	July	27	1	0	2	2	0.0	0	BB	GBR	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	98.00	0	1	2.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
119385	City Hotel	0	23	2017	August	35	30	2	5	2	0.0	0	BB	BEL	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	96.14	0	0	2.0
119386	City Hotel	0	102	2017	August	35	31	2	5	3	0.0	0	BB	FRA	TA/TO	0	0	0	E	E	0	No Deposit	0	Transient	225.43	0	2	3.0
119387	City Hotel	0	34	2017	August	35	31	2	5	2	0.0	0	BB	DEU	TA/TO	0	0	0	D	D	0	No Deposit	0	Transient	157.71	0	4	2.0
119388	City Hotel	0	109	2017	August	35	31	2	5	2	0.0	0	BB	GBR	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	104.40	0	0	2.0
119389	City Hotel	0	205	2017	August	35	29	2	7	2	0.0	0	HB	DEU	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	151.20	0	2	2.0

118728 rows × 28 columns

In [ ]:

hotel_df['total_nights'] = hotel_df['stays_in_week_nights'] + hotel_df['stays_in_weekend_nights']

<ipython-input-35-9d363fd31157>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df['total_nights'] = hotel_df['stays_in_week_nights'] + hotel_df['stays_in_weekend_nights']

In [ ]:

hotel_df[hotel_df['total_nights'] == 0]

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	people	total_nights
0	Resort Hotel	0	342	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	3	No Deposit	0	Transient	0.0	0	0	2.0	0
1	Resort Hotel	0	737	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	4	No Deposit	0	Transient	0.0	0	0	2.0	0
167	Resort Hotel	0	111	2015	July	28	6	0	0	2	0.0	0	BB	PRT	TA/TO	0	0	0	A	H	0	No Deposit	0	Transient	0.0	0	2	2.0	0
168	Resort Hotel	0	0	2015	July	28	6	0	0	1	0.0	0	BB	PRT	Direct	0	0	0	E	H	0	No Deposit	0	Transient	0.0	0	0	1.0	0
196	Resort Hotel	0	8	2015	July	28	7	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	A	A	0	No Deposit	0	Transient	0.0	0	1	2.0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
115483	City Hotel	0	15	2017	July	27	6	0	0	1	0.0	0	SC	FRA	Direct	0	0	0	A	A	0	No Deposit	0	Transient-Party	0.0	0	0	1.0	0
117701	City Hotel	0	0	2017	August	32	8	0	0	2	0.0	0	BB	PRT	TA/TO	1	0	0	A	A	0	No Deposit	0	Transient	0.0	0	0	2.0	0
118029	City Hotel	0	0	2017	August	33	14	0	0	2	0.0	0	BB	PRT	Direct	1	0	0	A	A	0	No Deposit	0	Transient	0.0	0	0	2.0	0
118631	City Hotel	0	78	2017	August	34	23	0	0	1	0.0	0	BB	PRT	TA/TO	0	0	0	A	K	7	No Deposit	0	Transient-Party	0.0	0	0	1.0	0
118963	City Hotel	0	1	2017	August	35	27	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	A	A	0	No Deposit	0	Transient	0.0	0	0	2.0	0

640 rows × 29 columns

In [ ]:

hotel_df['arrival_date_month'].apply(lambda x: 'spring' if x in ['March', 'April', 'May'] else 'summer' if x in ['June', 'July', 'August'] else 'fall' if x in ['September', 'October', 'November'] else 'winter')

Out[ ]:

0         summer
1         summer
2         summer
3         summer
4         summer
           ...  
119385    summer
119386    summer
119387    summer
119388    summer
119389    summer
Name: arrival_date_month, Length: 118728, dtype: object

In [ ]:

season_dic = {'spring':[3, 4, 5], 'summer':[6, 7, 8], 'fall':[9, 10, 11], 'winter':[12, 1, 2]}

In [ ]:

new_season_dic = {}

for i in season_dic:
    for j in season_dic[i]:
        new_season_dic[calendar.month_name[j]] = i

In [ ]:

new_season_dic

Out[ ]:

{'March': 'spring',
 'April': 'spring',
 'May': 'spring',
 'June': 'summer',
 'July': 'summer',
 'August': 'summer',
 'September': 'fall',
 'October': 'fall',
 'November': 'fall',
 'December': 'winter',
 'January': 'winter',
 'February': 'winter'}

In [ ]:

hotel_df['season'] = hotel_df['arrival_date_month'].map(new_season_dic)

<ipython-input-41-71c40adf4d6e>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df['season'] = hotel_df['arrival_date_month'].map(new_season_dic)

In [ ]:

hotel_df.head()

Out[ ]:

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	meal	country	distribution_channel	reserved_room_type	assigned_room_type	booking_changes	deposit_type	customer_type	adr	total_of_special_requests	people	total_nights	season
0	Resort Hotel	342	2015	July	27	1	0	2	BB	PRT	Direct	C	C	3	No Deposit	Transient	0.0	0	2.0	0	summer
1	Resort Hotel	737	2015	July	27	1	0	2	BB	PRT	Direct	C	C	4	No Deposit	Transient	0.0	0	2.0	0	summer
2	Resort Hotel	7	2015	July	27	1	1	1	BB	GBR	Direct	A	C	0	No Deposit	Transient	75.0	0	1.0	1	summer
3	Resort Hotel	13	2015	July	27	1	1	1	BB	GBR	Corporate	A	A	0	No Deposit	Transient	75.0	0	1.0	1	summer
4	Resort Hotel	14	2015	July	27	1	2	2	BB	GBR	TA/TO	A	A	0	No Deposit	Transient	98.0	1	2.0	2	summer

In [ ]:

hotel_df['expected_room_type'] = (hotel_df['reserved_room_type'] == hotel_df['assigned_room_type']).astype(int)

<ipython-input-43-d7dcc6110bd8>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df['expected_room_type'] = (hotel_df['reserved_room_type'] == hotel_df['assigned_room_type']).astype(int)

In [ ]:

hotel_df.head()

Out[ ]:

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	meal	country	distribution_channel	reserved_room_type	assigned_room_type	booking_changes	deposit_type	customer_type	adr	total_of_special_requests	people	total_nights	season	expected_room_type
0	Resort Hotel	342	2015	July	27	1	0	2	BB	PRT	Direct	C	C	3	No Deposit	Transient	0.0	0	2.0	0	summer	1
1	Resort Hotel	737	2015	July	27	1	0	2	BB	PRT	Direct	C	C	4	No Deposit	Transient	0.0	0	2.0	0	summer	1
2	Resort Hotel	7	2015	July	27	1	1	1	BB	GBR	Direct	A	C	0	No Deposit	Transient	75.0	0	1.0	1	summer	0
3	Resort Hotel	13	2015	July	27	1	1	1	BB	GBR	Corporate	A	A	0	No Deposit	Transient	75.0	0	1.0	1	summer	1
4	Resort Hotel	14	2015	July	27	1	2	2	BB	GBR	TA/TO	A	A	0	No Deposit	Transient	98.0	1	2.0	2	summer	1

In [ ]:

hotel_df['cancel_rate'] = hotel_df['previous_cancellations'] / (hotel_df['previous_cancellations'] + hotel_df['previous_bookings_not_canceled'])

<ipython-input-45-01c32095fb5f>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df['cancel_rate'] = hotel_df['previous_cancellations'] / (hotel_df['previous_cancellations'] + hotel_df['previous_bookings_not_canceled'])

In [ ]:

hotel_df.head()

Out[ ]:

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	meal	country	distribution_channel	reserved_room_type	assigned_room_type	booking_changes	deposit_type	customer_type	adr	total_of_special_requests	people	total_nights	season	expected_room_type	cancel_rate
0	Resort Hotel	342	2015	July	27	1	0	2	BB	PRT	Direct	C	C	3	No Deposit	Transient	0.0	0	2.0	0	summer	1	NaN
1	Resort Hotel	737	2015	July	27	1	0	2	BB	PRT	Direct	C	C	4	No Deposit	Transient	0.0	0	2.0	0	summer	1	NaN
2	Resort Hotel	7	2015	July	27	1	1	1	BB	GBR	Direct	A	C	0	No Deposit	Transient	75.0	0	1.0	1	summer	0	NaN
3	Resort Hotel	13	2015	July	27	1	1	1	BB	GBR	Corporate	A	A	0	No Deposit	Transient	75.0	0	1.0	1	summer	1	NaN
4	Resort Hotel	14	2015	July	27	1	2	2	BB	GBR	TA/TO	A	A	0	No Deposit	Transient	98.0	1	2.0	2	summer	1	NaN

In [ ]:

hotel_df[hotel_df['cancel_rate'].isna()] # 처음 방문한 사람들

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	people	total_nights	season	expected_room_type	cancel_rate
0	Resort Hotel	0	342	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	3	No Deposit	0	Transient	0.00	0	0	2.0	0	summer	1	NaN
1	Resort Hotel	0	737	2015	July	27	1	0	0	2	0.0	0	BB	PRT	Direct	0	0	0	C	C	4	No Deposit	0	Transient	0.00	0	0	2.0	0	summer	1	NaN
2	Resort Hotel	0	7	2015	July	27	1	0	1	1	0.0	0	BB	GBR	Direct	0	0	0	A	C	0	No Deposit	0	Transient	75.00	0	0	1.0	1	summer	0	NaN
3	Resort Hotel	0	13	2015	July	27	1	0	1	1	0.0	0	BB	GBR	Corporate	0	0	0	A	A	0	No Deposit	0	Transient	75.00	0	0	1.0	1	summer	1	NaN
4	Resort Hotel	0	14	2015	July	27	1	0	2	2	0.0	0	BB	GBR	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	98.00	0	1	2.0	2	summer	1	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
119385	City Hotel	0	23	2017	August	35	30	2	5	2	0.0	0	BB	BEL	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	96.14	0	0	2.0	7	summer	1	NaN
119386	City Hotel	0	102	2017	August	35	31	2	5	3	0.0	0	BB	FRA	TA/TO	0	0	0	E	E	0	No Deposit	0	Transient	225.43	0	2	3.0	7	summer	1	NaN
119387	City Hotel	0	34	2017	August	35	31	2	5	2	0.0	0	BB	DEU	TA/TO	0	0	0	D	D	0	No Deposit	0	Transient	157.71	0	4	2.0	7	summer	1	NaN
119388	City Hotel	0	109	2017	August	35	31	2	5	2	0.0	0	BB	GBR	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	104.40	0	0	2.0	7	summer	1	NaN
119389	City Hotel	0	205	2017	August	35	29	2	7	2	0.0	0	HB	DEU	TA/TO	0	0	0	A	A	0	No Deposit	0	Transient	151.20	0	2	2.0	9	summer	1	NaN

109523 rows × 32 columns

In [ ]:

hotel_df[~hotel_df['cancel_rate'].isna()] # 방문한적은 있지만 캔슬이 없는 경우

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	people	total_nights	season	expected_room_type	cancel_rate
13808	Resort Hotel	0	6	2016	January	5	26	0	2	1	0.0	0	BB	PRT	Corporate	0	0	1	A	D	1	No Deposit	0	Transient	27.0	0	0	1.0	2	winter	0	0.0
13813	Resort Hotel	0	1	2016	February	6	2	0	2	1	0.0	0	BB	PRT	Corporate	0	0	1	A	D	1	No Deposit	0	Transient	27.0	0	0	1.0	2	winter	0	0.0
13814	Resort Hotel	0	6	2016	November	47	14	1	0	1	0.0	0	BB	PRT	Corporate	0	0	2	A	A	0	No Deposit	0	Transient	27.0	0	0	1.0	1	fall	1	0.0
13815	Resort Hotel	0	6	2017	January	3	17	0	1	1	0.0	0	BB	PRT	Corporate	0	0	3	A	A	0	No Deposit	0	Transient	35.0	0	0	1.0	1	winter	1	0.0
13817	Resort Hotel	0	1	2017	February	8	21	0	2	1	0.0	0	BB	PRT	Corporate	0	0	1	A	A	0	No Deposit	0	Transient-Party	35.0	0	0	1.0	2	winter	1	0.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
117424	City Hotel	0	3	2017	August	35	31	0	1	2	1.0	0	BB	PRT	Corporate	1	0	1	A	A	0	No Deposit	0	Transient	95.0	0	4	3.0	1	summer	1	0.0
117841	City Hotel	0	7	2017	August	35	30	0	2	1	0.0	0	BB	PRT	Corporate	1	0	1	A	A	0	No Deposit	0	Transient	65.0	0	2	1.0	2	summer	1	0.0
118581	City Hotel	0	11	2017	August	34	25	0	2	2	0.0	0	BB	FRA	TA/TO	0	0	1	D	D	1	No Deposit	0	Group	125.0	0	0	2.0	2	summer	1	0.0
118651	City Hotel	0	189	2017	August	35	27	2	0	2	0.0	0	BB	ITA	TA/TO	0	0	1	A	A	1	No Deposit	0	Transient-Party	119.0	0	3	2.0	2	summer	1	0.0
118654	City Hotel	0	189	2017	August	35	27	2	0	2	0.0	0	BB	ITA	TA/TO	0	0	1	A	A	1	No Deposit	0	Transient	119.0	0	2	2.0	2	summer	1	0.0

9205 rows × 32 columns

In [ ]:

hotel_df[hotel_df['cancel_rate'] > 0] # 취소한 경우

Out[ ]:

	hotel	is_canceled	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	meal	country	distribution_channel	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	reserved_room_type	assigned_room_type	booking_changes	deposit_type	days_in_waiting_list	customer_type	adr	required_car_parking_spaces	total_of_special_requests	people	total_nights	season	expected_room_type	cancel_rate
13825	Resort Hotel	0	6	2016	March	13	21	1	0	1	0.0	0	BB	PRT	Corporate	0	1	1	A	A	0	No Deposit	0	Transient	40.0	0	0	1.0	1	spring	1	0.500000
13826	Resort Hotel	0	7	2016	June	26	21	0	1	1	0.0	0	BB	PRT	Corporate	0	1	2	A	A	0	No Deposit	0	Transient	65.0	0	0	1.0	1	summer	1	0.333333
13827	Resort Hotel	0	8	2016	September	40	27	0	2	2	0.0	0	BB	PRT	Corporate	0	1	3	A	A	0	No Deposit	0	Transient	65.0	0	0	2.0	2	fall	1	0.250000
13855	Resort Hotel	0	5	2015	November	48	25	0	1	1	0.0	0	BB	PRT	Corporate	0	1	1	A	A	0	No Deposit	0	Transient	25.0	0	0	1.0	1	fall	1	0.500000
13856	Resort Hotel	0	0	2015	December	52	22	0	1	1	0.0	0	BB	PRT	Corporate	0	1	2	A	A	0	No Deposit	0	Transient	25.0	0	0	1.0	1	winter	1	0.333333
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
111356	City Hotel	0	10	2017	June	25	22	0	1	1	0.0	0	BB	PRT	Corporate	1	1	4	A	A	0	No Deposit	0	Transient	65.0	0	0	1.0	1	summer	1	0.200000
111357	City Hotel	0	20	2017	July	28	11	0	3	1	0.0	0	BB	PRT	Corporate	1	1	5	A	A	0	No Deposit	0	Transient	65.0	0	0	1.0	3	summer	1	0.166667
111358	City Hotel	0	8	2017	July	30	25	0	1	1	0.0	0	BB	PRT	Corporate	1	1	6	A	A	0	No Deposit	0	Transient	65.0	1	0	1.0	1	summer	1	0.142857
111359	City Hotel	0	13	2017	August	35	29	0	1	1	0.0	0	BB	PRT	Corporate	1	1	7	A	A	0	No Deposit	0	Transient	65.0	0	0	1.0	1	summer	1	0.125000
111925	City Hotel	1	6	2017	July	29	17	1	0	1	0.0	0	BB	PRT	Corporate	1	1	1	A	D	0	No Deposit	0	Transient	65.0	0	0	1.0	1	summer	0	0.500000

6442 rows × 32 columns

In [ ]:

hotel_df['cancel_rate'] = hotel_df['cancel_rate'].fillna(-1)

<ipython-input-50-599e806aa6a0>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df['cancel_rate'] = hotel_df['cancel_rate'].fillna(-1)

In [ ]:

hotel_df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 118728 entries, 0 to 119389
Data columns (total 32 columns):
 #   Column                          Non-Null Count   Dtype  
---  ------                          --------------   -----  
 0   hotel                           118728 non-null  object 
 1   is_canceled                     118728 non-null  int64  
 2   lead_time                       118728 non-null  int64  
 3   arrival_date_year               118728 non-null  int64  
 4   arrival_date_month              118728 non-null  object 
 5   arrival_date_week_number        118728 non-null  int64  
 6   arrival_date_day_of_month       118728 non-null  int64  
 7   stays_in_weekend_nights         118728 non-null  int64  
 8   stays_in_week_nights            118728 non-null  int64  
 9   adults                          118728 non-null  int64  
 10  children                        118728 non-null  float64
 11  babies                          118728 non-null  int64  
 12  meal                            118728 non-null  object 
 13  country                         118728 non-null  object 
 14  distribution_channel            118728 non-null  object 
 15  is_repeated_guest               118728 non-null  int64  
 16  previous_cancellations          118728 non-null  int64  
 17  previous_bookings_not_canceled  118728 non-null  int64  
 18  reserved_room_type              118728 non-null  object 
 19  assigned_room_type              118728 non-null  object 
 20  booking_changes                 118728 non-null  int64  
 21  deposit_type                    118728 non-null  object 
 22  days_in_waiting_list            118728 non-null  int64  
 23  customer_type                   118728 non-null  object 
 24  adr                             118728 non-null  float64
 25  required_car_parking_spaces     118728 non-null  int64  
 26  total_of_special_requests       118728 non-null  int64  
 27  people                          118728 non-null  float64
 28  total_nights                    118728 non-null  int64  
 29  season                          118728 non-null  object 
 30  expected_room_type              118728 non-null  int64  
 31  cancel_rate                     118728 non-null  float64
dtypes: float64(4), int64(18), object(10)
memory usage: 29.9+ MB

In [ ]:

hotel_df['country'].dtype # dtype('O')

Out[ ]:

dtype('O')

In [ ]:

hotel_df['people'].dtype # dtype('float64')

Out[ ]:

dtype('float64')

In [ ]:

obj_list = []

for i in hotel_df.columns:
    if hotel_df[i].dtype == 'O':
        obj_list.append(i)

In [ ]:

obj_list

Out[ ]:

['hotel',
 'arrival_date_month',
 'meal',
 'country',
 'distribution_channel',
 'reserved_room_type',
 'assigned_room_type',
 'deposit_type',
 'customer_type',
 'season']

In [ ]:

for i in obj_list:
    print(i, hotel_df[i].nunique)

hotel <bound method IndexOpsMixin.nunique of 0         Resort Hotel
1         Resort Hotel
2         Resort Hotel
3         Resort Hotel
4         Resort Hotel
              ...     
119385      City Hotel
119386      City Hotel
119387      City Hotel
119388      City Hotel
119389      City Hotel
Name: hotel, Length: 118728, dtype: object>
arrival_date_month <bound method IndexOpsMixin.nunique of 0           July
1           July
2           July
3           July
4           July
           ...  
119385    August
119386    August
119387    August
119388    August
119389    August
Name: arrival_date_month, Length: 118728, dtype: object>
meal <bound method IndexOpsMixin.nunique of 0         BB
1         BB
2         BB
3         BB
4         BB
          ..
119385    BB
119386    BB
119387    BB
119388    BB
119389    HB
Name: meal, Length: 118728, dtype: object>
country <bound method IndexOpsMixin.nunique of 0         PRT
1         PRT
2         GBR
3         GBR
4         GBR
         ... 
119385    BEL
119386    FRA
119387    DEU
119388    GBR
119389    DEU
Name: country, Length: 118728, dtype: object>
distribution_channel <bound method IndexOpsMixin.nunique of 0            Direct
1            Direct
2            Direct
3         Corporate
4             TA/TO
            ...    
119385        TA/TO
119386        TA/TO
119387        TA/TO
119388        TA/TO
119389        TA/TO
Name: distribution_channel, Length: 118728, dtype: object>
reserved_room_type <bound method IndexOpsMixin.nunique of 0         C
1         C
2         A
3         A
4         A
         ..
119385    A
119386    E
119387    D
119388    A
119389    A
Name: reserved_room_type, Length: 118728, dtype: object>
assigned_room_type <bound method IndexOpsMixin.nunique of 0         C
1         C
2         C
3         A
4         A
         ..
119385    A
119386    E
119387    D
119388    A
119389    A
Name: assigned_room_type, Length: 118728, dtype: object>
deposit_type <bound method IndexOpsMixin.nunique of 0         No Deposit
1         No Deposit
2         No Deposit
3         No Deposit
4         No Deposit
             ...    
119385    No Deposit
119386    No Deposit
119387    No Deposit
119388    No Deposit
119389    No Deposit
Name: deposit_type, Length: 118728, dtype: object>
customer_type <bound method IndexOpsMixin.nunique of 0         Transient
1         Transient
2         Transient
3         Transient
4         Transient
            ...    
119385    Transient
119386    Transient
119387    Transient
119388    Transient
119389    Transient
Name: customer_type, Length: 118728, dtype: object>
season <bound method IndexOpsMixin.nunique of 0         summer
1         summer
2         summer
3         summer
4         summer
           ...  
119385    summer
119386    summer
119387    summer
119388    summer
119389    summer
Name: season, Length: 118728, dtype: object>

In [ ]:

hotel_df['meal'].value_counts()

Out[ ]:

BB           91789
HB           14429
SC           10547
Undefined     1165
FB             798
Name: meal, dtype: int64

In [ ]:

hotel_df.drop(['country', 'meal'], axis=1, inplace=True)

<ipython-input-58-55d6171d5808>:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  hotel_df.drop(['country', 'meal'], axis=1, inplace=True)

In [ ]:

obj_list.remove('country')
obj_list.remove('meal')

In [ ]:

hotel_df = pd.get_dummies(hotel_df, columns=obj_list)

In [ ]:

hotel_df.head()

Out[ ]:

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	booking_changes	adr	total_of_special_requests	people	total_nights	expected_room_type	cancel_rate	hotel_Resort Hotel	arrival_date_month_July	distribution_channel_Corporate	distribution_channel_Direct	distribution_channel_TA/TO	reserved_room_type_A	reserved_room_type_C	assigned_room_type_A	assigned_room_type_C	deposit_type_No Deposit	customer_type_Transient	season_summer
0	342	2015	27	1	0	2	3	0.0	0	2.0	0	1	-1.0	1	1	0	1	0	0	1	0	1	1	1	1
1	737	2015	27	1	0	2	4	0.0	0	2.0	0	1	-1.0	1	1	0	1	0	0	1	0	1	1	1	1
2	7	2015	27	1	1	1	0	75.0	0	1.0	1	0	-1.0	1	1	0	1	0	1	0	0	1	1	1	1
3	13	2015	27	1	1	1	0	75.0	0	1.0	1	1	-1.0	1	1	1	0	0	1	0	1	0	1	1	1
4	14	2015	27	1	2	2	0	98.0	1	2.0	2	1	-1.0	1	1	0	0	1	1	0	1	0	1	1	1

In [ ]:

from sklearn.model_selection import train_test_split

In [ ]:

X_train, X_test, y_train, y_test = train_test_split(hotel_df.drop('is_canceled', axis=1), hotel_df['is_canceled'], test_size=0.3, random_state=10)

2. 랜덤 포레스트(Random Forest)

Decision Tree는 매우 훌륭한 모델이지만 학습 데이터에 오버피팅 하는 경향이 있음(가지치기 같은 방법을 통해 부작용을 최소화할 수 있지만 부족함)
학습을 통해 구성해 놓은 다수의 나무들(Decision Tree)로부터 분류 결과를 취합해서 결론을 얻는 방식의 모델
Decision Tree 기반의 Bagging 앙상블 모델
굉장히 인기있는 모델이며, 사용성이 쉽고 성능도 꽤 우수한 편

2-1. 앙상블(Ensemble) 모델

여러개의 머신러닝 모델을 이용해 최적의 답을 찾아내는 기법
보팅(Voting)
- 다른 알고리즘 model을 조합해서 사용
- 모델에 대해 투표로 결과를 도출
배깅(Bagging)
- 같은 알고리즘 내에서 다른 sample 조합을 사
- 샘플 중복 생성을 통해 결과를 도출
부스팅(Boosting)
- 이전 오차를 보완해가면서 가중치를 부여
- 성능이 매우 우수하지만, 잘못된 레이블이나 이웃라이어에 대해 필요이상으로 민
스태킹(Stacking)
- 여러 모델을 기반으로 예측된 결과를 통해 meta 모델이 다시 한번 예측
- 성능을 극으로 끌어올릴 때 활용하지만 과대적합을 유발할 수 있음(특히 데이터셋이 적은 경우)

In [ ]:

from sklearn.ensemble import RandomForestClassifier

In [ ]:

rf = RandomForestClassifier()

In [ ]:

rf.fit(X_train, y_train)

In [ ]:

pred1 = rf.predict(X_test)

In [ ]:

pred1

Out[ ]:

array([0, 0, 0, ..., 0, 0, 0])

In [ ]:

proba1 = rf.predict_proba(X_test)
proba1

Out[ ]:

array([[0.955, 0.045],
       [1.   , 0.   ],
       [0.72 , 0.28 ],
       ...,
       [1.   , 0.   ],
       [0.9  , 0.1  ],
       [0.96 , 0.04 ]])

In [ ]:

# 첫번째 테스트 데이터에 대한 예측 결과
proba1[0]

Out[ ]:

array([0.955, 0.045])

In [ ]:

# 모든 테스트 데이터에 대한 호텔 예약을 취소할 확률만 출력
proba1[:, 1]

Out[ ]:

array([0.045, 0.   , 0.28 , ..., 0.   , 0.1  , 0.04 ])

3. ROC Curve

이진 분류의 성능을 측정하는 도구
민감도와 특이도를 그려지는 곡선을 의미
FPR(False Positive Rate)
- 특이도
- FP / TN + FP
- 실제값은 음성이지만 양성으로 잘 못 분류
TPR(True Positive Rate)
- 민감도. 참인 양성 비율
- TP / FN + TP
- 실제로 양성이고 양성으로 잘 분류

4. AUC(Area Under the ROC Curve)

ROC 커브와 직선 사이의 면적을 의미
AUC값의 범위는 0.5~1이며 값이 클수록 예측의 정확도가 높음

In [ ]:

from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, roc_auc_score

In [ ]:

accuracy_score(y_test, pred1)

Out[ ]:

0.8630225441477863

In [ ]:

confusion_matrix(y_test, pred1)

Out[ ]:

array([[20779,  1579],
       [ 3300,  9961]])

In [ ]:

print(classification_report(y_test, pred1))

              precision    recall  f1-score   support

           0       0.86      0.93      0.89     22358
           1       0.86      0.75      0.80     13261

    accuracy                           0.86     35619
   macro avg       0.86      0.84      0.85     35619
weighted avg       0.86      0.86      0.86     35619

In [ ]:

roc_auc_score(y_test, proba1[:, 1])

Out[ ]:

0.9303955154719541

In [ ]:

# 하이퍼 파라미터 수정
rf2 = RandomForestClassifier(max_depth=30, random_state=10)
rf2.fit(X_train, y_train)
proba2 = rf2.predict_proba(X_test)
roc_auc_score(y_test, proba2[:, 1])

# 하이퍼 파라미터 적용 전: 0.9303955154719541
# 하이퍼 파라미터 적용(max_depth=30 을 적용) 후: 0.9320285483491656

Out[ ]:

0.9320285483491656

In [ ]:

# 하이퍼 파라미터 수정(max_septh=50 을 적용)
rf2 = RandomForestClassifier(max_depth=50, random_state=10)
rf2.fit(X_train, y_train)
proba2 = rf2.predict_proba(X_test)
roc_auc_score(y_test, proba2[:, 1])

# 하이퍼 파라미터 적용 전: 0.9303955154719541
# 하이퍼 파라미터 적용(max_depth=30 을 적용) 후: 0.9320285483491656
# 하이퍼 파라미터 적용(max_depth=50 을 적용) 후: 0.9303745518246758

Out[ ]:

0.9303745518246758

In [ ]:

# 하이퍼 파라미터 수정(min_samples_split=5 을 적용)
rf2 = RandomForestClassifier(min_samples_split=5, random_state=10)
rf2.fit(X_train, y_train)
proba2 = rf2.predict_proba(X_test)
roc_auc_score(y_test, proba2[:, 1])

# 하이퍼 파라미터 적용 전: 0.9303955154719541
# 하이퍼 파라미터 적용(min_samples_split=5 을 적용) 후: 0.931436154565479

Out[ ]:

0.931436154565479

In [ ]:

# 하이퍼 파라미터 수정(min_samples_split=7 을 적용)
rf2 = RandomForestClassifier(min_samples_split=7, random_state=10)
rf2.fit(X_train, y_train)
proba2 = rf2.predict_proba(X_test)
roc_auc_score(y_test, proba2[:, 1])

# 하이퍼 파라미터 적용 전: 0.9303955154719541
# 하이퍼 파라미터 적용(min_samples_split=5 을 적용) 후: 0.931436154565479
# 하이퍼 파라미터 적용(min_samples_split=7 을 적용) 후: 0.9312578210627522

Out[ ]:

0.9312578210627522

In [ ]:

# 하이퍼 파라미터 최적의 값
rf2 = RandomForestClassifier(max_depth=30, min_samples_split=5, random_state=10)
rf2.fit(X_train, y_train)
proba2 = rf2.predict_proba(X_test)
roc_auc_score(y_test, proba2[:, 1])

# 하이퍼 파라미터 적용 전: 0.9303955154719541
# 하이퍼 파라미터 적용(max_depth=30 을 적용) 후: 0.9320285483491656
# 하이퍼 파라미터 적용(min_samples_split=5 을 적용) 후: 0.931436154565479
# 하이퍼 파라미터 적용(max_depth=30, min_samples_split=5) 후: 0.9320878540030826

Out[ ]:

0.9320878540030826

5. 하이퍼 파라미터 최적의 값을 찾는 방법

GridSearchCV: 원하는 모든 하이퍼 파라미터를 적용하여 최적의 값을 찾음

In [ ]:

from sklearn.model_selection import GridSearchCV

In [ ]:

params = {
    'max_depth': [None, 10, 30, 50],
    'min_samples_split': [2, 3, 5, 7, 10]
}

In [ ]:

rf3 = RandomForestClassifier(random_state=10)

In [ ]:

grid_df = GridSearchCV(rf3, params, cv=5) # cv: 데이터 교차검증

In [ ]:

grid_df.fit(X_train, y_train)

GridSearchCV

GridSearchCV(cv=5, estimator=RandomForestClassifier(random_state=10),
             param_grid={'max_depth': [None, 10, 30, 50],
                         'min_samples_split': [2, 3, 5, 7, 10]})

estimator: RandomForestClassifier

RandomForestClassifier(random_state=10)

RandomForestClassifier

RandomForestClassifier(random_state=10)

In [ ]:

grid_df.best_params_

Out[ ]:

{'max_depth': 30, 'min_samples_split': 2}

In [ ]:

rf3 = RandomForestClassifier(max_depth=30, min_samples_split=2, random_state=10)
rf3.fit(X_train, y_train)
proba3 = rf3.predict_proba(X_test)
roc_auc_score(y_test, proba3[:, 1])

# 하이퍼 파라미터 적용(max_depth=30, min_samples_split=5) 후: 0.9320285483491656

Out[ ]:

0.9320285483491656

6. 피쳐 중요도(Feature Importances)

Decisiom Tree에서 노드를 분기할 때 해당 피쳐가 클래스를 나누는데 얼마나 영향을 미쳤는지를 표기하는 척도
0이면 클래스를 구분하는데 해당 피쳐가 선택되지 않았다는 것이며, 1이면 해당 피쳐가 클래스를 완벽하게 나눴다는 것을 의미

In [ ]:

proba3 = rf3.predict_proba(X_test)
proba3

Out[ ]:

array([[0.96333333, 0.03666667],
       [0.98747831, 0.01252169],
       [0.67944316, 0.32055684],
       ...,
       [1.        , 0.        ],
       [0.90383261, 0.09616739],
       [0.97      , 0.03      ]])

In [ ]:

roc_auc_score(y_test, proba3[:, 1])

Out[ ]:

0.9320285483491656

In [ ]:

rf3.feature_importances_

Out[ ]:

array([1.27137138e-01, 2.14563953e-02, 4.68663037e-02, 5.96305446e-02,
       2.16746957e-02, 3.25152451e-02, 1.02347526e-02, 5.40307416e-03,
       8.84445875e-04, 2.02141842e-03, 2.44460558e-02, 3.14871409e-03,
       2.18776128e-02, 3.06142123e-03, 9.51803193e-02, 2.23244236e-02,
       5.75968205e-02, 1.36759899e-02, 3.53588232e-02, 2.80447856e-02,
       3.80039740e-02, 7.62870375e-03, 6.72543460e-03, 3.44290597e-03,
       4.03857944e-03, 2.04861894e-03, 2.63415501e-03, 1.84105707e-03,
       3.95932072e-03, 3.62772728e-03, 3.05844877e-03, 3.56107858e-03,
       2.38180451e-03, 3.26791942e-03, 2.91845948e-03, 3.01678974e-03,
       9.42477811e-03, 2.46969262e-04, 1.08307976e-02, 0.00000000e+00,
       5.74350945e-03, 9.03907596e-04, 6.54901146e-04, 4.05826969e-03,
       2.30412987e-03, 1.25861929e-03, 1.10018824e-03, 3.65382816e-04,
       3.43650984e-05, 1.10160554e-02, 1.51175712e-03, 1.19112472e-03,
       5.19366703e-03, 2.61859311e-03, 1.46746839e-03, 1.15967758e-03,
       3.98805744e-04, 1.35937049e-04, 1.43906056e-04, 2.28125116e-06,
       6.99747351e-02, 9.55569434e-02, 4.52910589e-04, 3.08796348e-03,
       4.79153629e-04, 1.67417487e-02, 1.19682713e-02, 3.45258733e-03,
       4.02311808e-03, 4.35533780e-03, 3.44818237e-03])

In [ ]:

1.27137138e-01

Out[ ]:

0.127137138

In [ ]:

feat_imp = pd.DataFrame({
    'features': X_train.columns,
    'importances': rf3.feature_importances_
})

In [ ]:

feat_imp

Out[ ]:

	features	importances
0	lead_time	0.127137
1	arrival_date_year	0.021456
2	arrival_date_week_number	0.046866
3	arrival_date_day_of_month	0.059631
4	stays_in_weekend_nights	0.021675
...	...	...
66	customer_type_Transient-Party	0.011968
67	season_fall	0.003453
68	season_spring	0.004023
69	season_summer	0.004355
70	season_winter	0.003448

71 rows × 2 columns

In [ ]:

top10 = feat_imp.sort_values('importances', ascending=False).head(10)
top10

Out[ ]:

	features	importances
0	lead_time	0.127137
61	deposit_type_Non Refund	0.095557
14	adr	0.095180
60	deposit_type_No Deposit	0.069975
3	arrival_date_day_of_month	0.059631
16	total_of_special_requests	0.057597
2	arrival_date_week_number	0.046866
20	cancel_rate	0.038004
18	total_nights	0.035359
5	stays_in_week_nights	0.032515

In [ ]:

plt.figure(figsize=(5, 10))
sns.barplot(x='importances', y='features', data=top10)

Out[ ]:

<Axes: xlabel='importances', ylabel='features'>

(Python) KMeans (1)	2023.06.15
(Python) lightGBM (0)	2023.06.15
(Python) 서포트 벡터 머신 (0)	2023.06.14
(Python) 로지스틱 회귀 (0)	2023.06.14
(Python) 의사 결정 나무 (0)	2023.06.14

내 블로그 - 관리자 홈 전환	`Q` `Q`
새 글 쓰기	`W` `W`

글 수정 (권한 있는 경우)	`E` `E`
댓글 영역으로 이동	`C` `C`

이 페이지의 URL 복사	`S` `S`
맨 위로 이동	`T` `T`
티스토리 홈 이동	`H` `H`
단축키 안내	`Shift` + `/` `⇧` + `/`

흰둥이는 코드를 짤 때 짖어 (왈!왈!왈!왈!왈!왈!왈!왈!왈!왈!왈!)

흰둥이는 코드를 짤 때 짖어 (왈!왈!왈!왈!왈!왈!왈!왈!왈!왈!왈!)

(Python) 랜덤 포레스트 본문

(Python) 랜덤 포레스트

1. hotel 데이터셋 살펴보기

2. 랜덤 포레스트(Random Forest)

2-1. 앙상블(Ensemble) 모델

3. ROC Curve

4. AUC(Area Under the ROC Curve)

5. 하이퍼 파라미터 최적의 값을 찾는 방법

6. 피쳐 중요도(Feature Importances)

'파이썬 머신러닝, 딥러닝' 카테고리의 다른 글

티스토리툴바

단축키

내 블로그

블로그 게시글

모든 영역

« 2025/04 »
일	월	화	수	목	금	토
		1	2	3	4	5
6	7	8	9	10	11	12
13	14	15	16	17	18	19
20	21	22	23	24	25	26
27	28	29	30