import os
import numpy as np # linear algebra
import pandas as pd
import os
for dirname, _, filenames in os.walk('/kaggle/input'):
for filename in filenames:
print(os.path.join(dirname, filename))
/kaggle/input/pubg-finish-placement-prediction/train_V2.csv /kaggle/input/pubg-finish-placement-prediction/test_V2.csv /kaggle/input/pubg-finish-placement-prediction/sample_submission_V2.csv
%%time
# Any results you write to the current directory are saved as output.
path = "/kaggle/input/pubg-finish-placement-prediction/"
df_train = pd.read_csv(path + "train_V2.csv")
df_train.shape
CPU times: user 20.9 s, sys: 3.04 s, total: 24 s Wall time: 32.4 s
(4446966, 29)
df_train.memory_usage().sum() / 1024**2
983.9022064208984
# iinfo() 는 ()안이 자료형이 표현가능한 한계를 반환한다.
ii16 = np.iinfo(np.int16)
print(ii16.min, ii16.max)
ii32 = np.iinfo(np.int32)
print(ii32.min, ii32.max)
-32768 32767 -2147483648 2147483647
def reduce_mem_usage(df):
start_mem = df.memory_usage().sum() / 1024**2
print('메모리 사용량 {:.2f} MB'.format(start_mem))
for col in df.columns:
col_type = df[col].dtype
if col_type != object:
c_min = df[col].min()
c_max = df[col].max()
if str(col_type)[:3] == 'int':
if c_min > np.iinfo(np.int8).min and c_max < np.iinfo(np.int8).max:
df[col] = df[col].astype(np.int8)
elif c_min > np.iinfo(np.int16).min and c_max < np.iinfo(np.int16).max:
df[col] = df[col].astype(np.int16)
elif c_min > np.iinfo(np.int32).min and c_max < np.iinfo(np.int32).max:
df[col] = df[col].astype(np.int32)
elif c_min > np.iinfo(np.int64).min and c_max < np.iinfo(np.int64).max:
df[col] = df[col].astype(np.int64)
else:
if c_min > np.finfo(np.float16).min and c_max < np.finfo(np.float16).max:
df[col] = df[col].astype(np.float16)
elif c_min > np.finfo(np.float32).min and c_max < np.finfo(np.float32).max:
df[col] = df[col].astype(np.float32)
else:
df[col] = df[col].astype(np.float64)
else:
df[col] = df[col].astype('category')
end_mem = df.memory_usage().sum() / 1024**2
print('Memory usage after optimization is: {:.2f} MB'.format(end_mem))
print('Decreased by {:.1f}%'.format(100 * (start_mem - end_mem) / start_mem))
return df
def import_data(file):
### 메모리를 최적화 시킨다.
df = pd.read_csv(file, parse_dates=True, keep_date_col=True)
df = reduce_mem_usage(df)
return df
%%time
path = "/kaggle/input/pubg-finish-placement-prediction/"
print('-' * 80)
print('train data loading...')
train = import_data(path + "train_V2.csv")
print('test data loading...')
test = import_data(path + "test_V2.csv")
print('submissiion csv loading...')
df_sub = import_data(path + "sample_submission_V2.csv")
train.shape, test.shape, df_sub.shape
-------------------------------------------------------------------------------- train data loading... 메모리 사용량 983.90 MB Memory usage after optimization is: 452.07 MB Decreased by 54.1% test data loading... 메모리 사용량 413.18 MB Memory usage after optimization is: 201.94 MB Decreased by 51.1% submissiion csv loading... 메모리 사용량 29.51 MB Memory usage after optimization is: 88.48 MB Decreased by -199.8% CPU times: user 1min 23s, sys: 3.44 s, total: 1min 26s Wall time: 1min 31s
((4446966, 29), (1934174, 28), (1934174, 2))
train.info(null_counts=True)
/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:1: FutureWarning: null_counts is deprecated. Use show_counts instead """Entry point for launching an IPython kernel.
<class 'pandas.core.frame.DataFrame'> RangeIndex: 4446966 entries, 0 to 4446965 Data columns (total 29 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Id 4446966 non-null category 1 groupId 4446966 non-null category 2 matchId 4446966 non-null category 3 assists 4446966 non-null int8 4 boosts 4446966 non-null int8 5 damageDealt 4446966 non-null float16 6 DBNOs 4446966 non-null int8 7 headshotKills 4446966 non-null int8 8 heals 4446966 non-null int8 9 killPlace 4446966 non-null int8 10 killPoints 4446966 non-null int16 11 kills 4446966 non-null int8 12 killStreaks 4446966 non-null int8 13 longestKill 4446966 non-null float16 14 matchDuration 4446966 non-null int16 15 matchType 4446966 non-null category 16 maxPlace 4446966 non-null int8 17 numGroups 4446966 non-null int8 18 rankPoints 4446966 non-null int16 19 revives 4446966 non-null int8 20 rideDistance 4446966 non-null float16 21 roadKills 4446966 non-null int8 22 swimDistance 4446966 non-null float16 23 teamKills 4446966 non-null int8 24 vehicleDestroys 4446966 non-null int8 25 walkDistance 4446966 non-null float16 26 weaponsAcquired 4446966 non-null int16 27 winPoints 4446966 non-null int16 28 winPlacePerc 4446965 non-null float16 dtypes: category(4), float16(6), int16(5), int8(14) memory usage: 452.1 MB
test.info(null_counts=True)
/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:1: FutureWarning: null_counts is deprecated. Use show_counts instead """Entry point for launching an IPython kernel.
<class 'pandas.core.frame.DataFrame'> RangeIndex: 1934174 entries, 0 to 1934173 Data columns (total 28 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Id 1934174 non-null category 1 groupId 1934174 non-null category 2 matchId 1934174 non-null category 3 assists 1934174 non-null int8 4 boosts 1934174 non-null int8 5 damageDealt 1934174 non-null float16 6 DBNOs 1934174 non-null int8 7 headshotKills 1934174 non-null int8 8 heals 1934174 non-null int8 9 killPlace 1934174 non-null int8 10 killPoints 1934174 non-null int16 11 kills 1934174 non-null int8 12 killStreaks 1934174 non-null int8 13 longestKill 1934174 non-null float16 14 matchDuration 1934174 non-null int16 15 matchType 1934174 non-null category 16 maxPlace 1934174 non-null int8 17 numGroups 1934174 non-null int8 18 rankPoints 1934174 non-null int16 19 revives 1934174 non-null int8 20 rideDistance 1934174 non-null float16 21 roadKills 1934174 non-null int8 22 swimDistance 1934174 non-null float16 23 teamKills 1934174 non-null int8 24 vehicleDestroys 1934174 non-null int8 25 walkDistance 1934174 non-null float16 26 weaponsAcquired 1934174 non-null int16 27 winPoints 1934174 non-null int16 dtypes: category(4), float16(5), int16(5), int8(14) memory usage: 201.9 MB
train.winPlacePerc.value_counts().head()
0.000000 220505 1.000000 127573 0.500000 55065 0.333252 42508 0.666504 38112 Name: winPlacePerc, dtype: int64
train.nunique()
Id 4446966 groupId 2026745 matchId 47965 assists 20 boosts 27 damageDealt 12925 DBNOs 39 headshotKills 34 heals 63 killPlace 101 killPoints 1707 kills 58 killStreaks 18 longestKill 11155 matchDuration 1267 matchType 16 maxPlace 100 numGroups 100 rankPoints 2262 revives 25 rideDistance 15734 roadKills 14 swimDistance 12589 teamKills 11 vehicleDestroys 6 walkDistance 15940 weaponsAcquired 97 winPoints 1447 winPlacePerc 2268 dtype: int64
train.dropna(axis=0,inplace=True)
train.info(null_counts=True)
/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: null_counts is deprecated. Use show_counts instead
<class 'pandas.core.frame.DataFrame'> Int64Index: 4446965 entries, 0 to 4446965 Data columns (total 29 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Id 4446965 non-null category 1 groupId 4446965 non-null category 2 matchId 4446965 non-null category 3 assists 4446965 non-null int8 4 boosts 4446965 non-null int8 5 damageDealt 4446965 non-null float16 6 DBNOs 4446965 non-null int8 7 headshotKills 4446965 non-null int8 8 heals 4446965 non-null int8 9 killPlace 4446965 non-null int8 10 killPoints 4446965 non-null int16 11 kills 4446965 non-null int8 12 killStreaks 4446965 non-null int8 13 longestKill 4446965 non-null float16 14 matchDuration 4446965 non-null int16 15 matchType 4446965 non-null category 16 maxPlace 4446965 non-null int8 17 numGroups 4446965 non-null int8 18 rankPoints 4446965 non-null int16 19 revives 4446965 non-null int8 20 rideDistance 4446965 non-null float16 21 roadKills 4446965 non-null int8 22 swimDistance 4446965 non-null float16 23 teamKills 4446965 non-null int8 24 vehicleDestroys 4446965 non-null int8 25 walkDistance 4446965 non-null float16 26 weaponsAcquired 4446965 non-null int16 27 winPoints 4446965 non-null int16 28 winPlacePerc 4446965 non-null float16 dtypes: category(4), float16(6), int16(5), int8(14) memory usage: 486.0 MB
train.corr()
| assists | boosts | damageDealt | DBNOs | headshotKills | heals | killPlace | killPoints | kills | killStreaks | ... | revives | rideDistance | roadKills | swimDistance | teamKills | vehicleDestroys | walkDistance | weaponsAcquired | winPoints | winPlacePerc | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| assists | 1.000000 | 0.307683 | 0.406726 | 0.301057 | 0.198289 | 0.228556 | -0.290062 | 0.039066 | 0.319690 | 0.243471 | ... | 0.198320 | 0.110644 | 0.011903 | 0.023372 | 0.006081 | 0.057921 | 0.290305 | 0.243882 | 0.024014 | 0.299439 |
| boosts | 0.307683 | 1.000000 | 0.521947 | 0.358907 | 0.334661 | 0.535854 | -0.554844 | 0.008347 | 0.502024 | 0.405327 | ... | 0.253125 | 0.328855 | 0.035124 | 0.107992 | 0.013068 | 0.087387 | 0.640150 | 0.406609 | -0.007645 | 0.634232 |
| damageDealt | 0.406726 | 0.521947 | 1.000000 | 0.735762 | 0.613409 | 0.342987 | -0.677398 | 0.049904 | 0.888784 | 0.703599 | ... | 0.256951 | 0.140883 | 0.052487 | 0.036729 | 0.015468 | 0.081182 | 0.398199 | 0.353249 | 0.017762 | 0.440506 |
| DBNOs | 0.301057 | 0.358907 | 0.735762 | 1.000000 | 0.469923 | 0.265485 | -0.555290 | 0.042616 | 0.707848 | 0.646872 | ... | 0.300999 | 0.102536 | 0.036093 | 0.017320 | 0.069430 | 0.060210 | 0.284660 | 0.218161 | 0.011108 | 0.279968 |
| headshotKills | 0.198289 | 0.334661 | 0.613409 | 0.469923 | 1.000000 | 0.199917 | -0.469829 | 0.023988 | 0.674275 | 0.512882 | ... | 0.150145 | 0.076494 | 0.013369 | 0.028531 | 0.008658 | 0.039200 | 0.250982 | 0.217552 | 0.004982 | 0.277722 |
| heals | 0.228556 | 0.535854 | 0.342987 | 0.265485 | 0.199917 | 1.000000 | -0.386032 | -0.002585 | 0.311781 | 0.270412 | ... | 0.236680 | 0.297484 | 0.024619 | 0.079586 | 0.035386 | 0.062870 | 0.430266 | 0.309071 | -0.010990 | 0.427856 |
| killPlace | -0.290062 | -0.554844 | -0.677398 | -0.555290 | -0.469829 | -0.386032 | 1.000000 | -0.023278 | -0.730499 | -0.810517 | ... | -0.269804 | -0.236202 | -0.057356 | -0.089396 | -0.037797 | -0.073569 | -0.590236 | -0.490144 | -0.002415 | -0.719069 |
| killPoints | 0.039066 | 0.008347 | 0.049904 | 0.042616 | 0.023988 | -0.002585 | -0.023278 | 1.000000 | 0.036192 | 0.037091 | ... | 0.012041 | -0.029316 | -0.004156 | -0.004604 | -0.039739 | -0.009658 | 0.004630 | -0.003978 | 0.983417 | 0.012909 |
| kills | 0.319690 | 0.502024 | 0.888784 | 0.707848 | 0.674275 | 0.311781 | -0.730499 | 0.036192 | 1.000000 | 0.802500 | ... | 0.253323 | 0.114749 | 0.063651 | 0.039127 | 0.017601 | 0.079991 | 0.374774 | 0.336414 | 0.006929 | 0.419915 |
| killStreaks | 0.243471 | 0.405327 | 0.703599 | 0.646872 | 0.512882 | 0.270412 | -0.810517 | 0.037091 | 0.802500 | 1.000000 | ... | 0.241695 | 0.098420 | 0.065575 | 0.034588 | 0.027612 | 0.064374 | 0.327639 | 0.275373 | 0.008552 | 0.377566 |
| longestKill | 0.261426 | 0.423291 | 0.563838 | 0.451422 | 0.447261 | 0.263278 | -0.541147 | 0.014084 | 0.602937 | 0.515097 | ... | 0.170134 | 0.190853 | 0.025214 | 0.053455 | 0.007604 | 0.083227 | 0.413810 | 0.276264 | -0.003720 | 0.410153 |
| matchDuration | -0.019450 | 0.072107 | -0.006755 | -0.014486 | -0.017657 | 0.108901 | -0.005806 | -0.099542 | -0.030688 | -0.034645 | ... | 0.016990 | 0.365212 | 0.020016 | 0.008737 | 0.047969 | 0.060924 | 0.145062 | 0.129080 | -0.087361 | -0.005171 |
| maxPlace | -0.147916 | -0.013686 | -0.040708 | -0.267710 | 0.009211 | -0.064759 | 0.016791 | -0.044802 | -0.004847 | -0.037339 | ... | -0.160369 | -0.047784 | 0.011504 | 0.007788 | -0.038057 | -0.018888 | -0.083624 | -0.039121 | -0.029295 | 0.037381 |
| numGroups | -0.146805 | -0.012929 | -0.040078 | -0.265789 | 0.009389 | -0.064204 | 0.018786 | -0.042367 | -0.004060 | -0.036054 | ... | -0.159173 | -0.049880 | 0.010435 | 0.007921 | -0.038429 | -0.019173 | -0.082667 | -0.039984 | -0.027483 | 0.039625 |
| rankPoints | -0.016407 | 0.023202 | -0.001459 | -0.003451 | 0.005119 | 0.019142 | -0.014805 | -0.975555 | 0.008200 | 0.003426 | ... | -0.000613 | 0.036813 | 0.006138 | 0.004724 | 0.039626 | 0.013400 | 0.015790 | 0.014756 | -0.993845 | 0.013522 |
| revives | 0.198320 | 0.253125 | 0.256951 | 0.300999 | 0.150145 | 0.236680 | -0.269804 | 0.012041 | 0.253323 | 0.241695 | ... | 1.000000 | 0.109946 | 0.005920 | 0.020338 | 0.032989 | 0.038019 | 0.236796 | 0.167281 | 0.002434 | 0.240880 |
| rideDistance | 0.110644 | 0.328855 | 0.140883 | 0.102536 | 0.076494 | 0.297484 | -0.236202 | -0.029316 | 0.114749 | 0.098420 | ... | 0.109946 | 1.000000 | 0.103518 | 0.049448 | 0.050178 | 0.119254 | 0.310936 | 0.282246 | -0.027809 | 0.342914 |
| roadKills | 0.011903 | 0.035124 | 0.052487 | 0.036093 | 0.013369 | 0.024619 | -0.057356 | -0.004156 | 0.063651 | 0.065575 | ... | 0.005920 | 0.103518 | 1.000000 | 0.004411 | 0.009684 | 0.031071 | 0.016736 | 0.016494 | -0.004750 | 0.034544 |
| swimDistance | 0.023372 | 0.107992 | 0.036729 | 0.017320 | 0.028531 | 0.079586 | -0.089396 | -0.004604 | 0.039127 | 0.034588 | ... | 0.020338 | 0.049448 | 0.004411 | 1.000000 | 0.006200 | 0.008090 | 0.167509 | 0.075698 | -0.002680 | 0.149606 |
| teamKills | 0.006081 | 0.013068 | 0.015468 | 0.069430 | 0.008658 | 0.035386 | -0.037797 | -0.039739 | 0.017601 | 0.027612 | ... | 0.032989 | 0.050178 | 0.009684 | 0.006200 | 1.000000 | 0.112164 | 0.024858 | 0.035918 | -0.040943 | 0.015942 |
| vehicleDestroys | 0.057921 | 0.087387 | 0.081182 | 0.060210 | 0.039200 | 0.062870 | -0.073569 | -0.009658 | 0.079991 | 0.064374 | ... | 0.038019 | 0.119254 | 0.031071 | 0.008090 | 0.112164 | 1.000000 | 0.076237 | 0.061012 | -0.011426 | 0.073436 |
| walkDistance | 0.290305 | 0.640150 | 0.398199 | 0.284660 | 0.250982 | 0.430266 | -0.590236 | 0.004630 | 0.374774 | 0.327639 | ... | 0.236796 | 0.310936 | 0.016736 | 0.167509 | 0.024858 | 0.076237 | 1.000000 | 0.541222 | -0.002043 | 0.810886 |
| weaponsAcquired | 0.243882 | 0.406609 | 0.353249 | 0.218161 | 0.217552 | 0.309071 | -0.490144 | -0.003978 | 0.336414 | 0.275373 | ... | 0.167281 | 0.282246 | 0.016494 | 0.075698 | 0.035918 | 0.061012 | 0.541222 | 1.000000 | -0.003815 | 0.583806 |
| winPoints | 0.024014 | -0.007645 | 0.017762 | 0.011108 | 0.004982 | -0.010990 | -0.002415 | 0.983417 | 0.006929 | 0.008552 | ... | 0.002434 | -0.027809 | -0.004750 | -0.002680 | -0.040943 | -0.011426 | -0.002043 | -0.003815 | 1.000000 | 0.007062 |
| winPlacePerc | 0.299439 | 0.634232 | 0.440506 | 0.279968 | 0.277722 | 0.427856 | -0.719069 | 0.012909 | 0.419915 | 0.377566 | ... | 0.240880 | 0.342914 | 0.034544 | 0.149606 | 0.015942 | 0.073436 | 0.810886 | 0.583806 | 0.007062 | 1.000000 |
25 rows × 25 columns
train.corr()['winPlacePerc'].sort_values()
killPlace -0.719069 matchDuration -0.005171 winPoints 0.007062 killPoints 0.012909 rankPoints 0.013522 teamKills 0.015942 roadKills 0.034544 maxPlace 0.037381 numGroups 0.039625 vehicleDestroys 0.073436 swimDistance 0.149606 revives 0.240880 headshotKills 0.277722 DBNOs 0.279968 assists 0.299439 rideDistance 0.342914 killStreaks 0.377566 longestKill 0.410153 kills 0.419915 heals 0.427856 damageDealt 0.440506 weaponsAcquired 0.583806 boosts 0.634232 walkDistance 0.810886 winPlacePerc 1.000000 Name: winPlacePerc, dtype: float64
### 가장 상관관계 높은 2개의 변수를 선택
train[ ['killPlace', 'walkDistance']].head()
| killPlace | walkDistance | |
|---|---|---|
| 0 | 60 | 244.75 |
| 1 | 57 | 1434.00 |
| 2 | 47 | 161.75 |
| 3 | 75 | 202.75 |
| 4 | 45 | 49.75 |
sel = ['killPlace', 'walkDistance']
X_tr = train[sel]
y_tr = train['winPlacePerc']
X_last_test = test[sel]
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
X_train, X_test, y_train, y_test = train_test_split(X_tr, y_tr, test_size=0.2, random_state=77)
X_train.shape, X_test.shape
((3557572, 2), (889393, 2))
lr = LinearRegression()
lr.fit(X_train, y_train)
print("학습용 score() - 결정계수 : ", lr.score(X_train, y_train))
print("테스트용 score() - 결정계수 : ", lr.score(X_test, y_test))
학습용 score() - 결정계수 : 0.7464619534577828 테스트용 score() - 결정계수 : 0.7455769968673939
%%time
cvs_lr = cross_val_score(lr, X_test, y_test, cv=5, scoring="neg_median_absolute_error")
cvs_lr.mean(), cvs_lr.std()
CPU times: user 1.35 s, sys: 641 ms, total: 1.99 s Wall time: 1.03 s
(-0.09620575904846192, 0.00020372508052395657)
df_sub.head()
| Id | winPlacePerc | |
|---|---|---|
| 0 | 9329eb41e215eb | 1 |
| 1 | 639bd0dcd7bda8 | 1 |
| 2 | 63d5c8ef8dfe91 | 1 |
| 3 | cf5b81422591d1 | 1 |
| 4 | ee6a295187ba21 | 1 |
test.head()
| Id | groupId | matchId | assists | boosts | damageDealt | DBNOs | headshotKills | heals | killPlace | ... | rankPoints | revives | rideDistance | roadKills | swimDistance | teamKills | vehicleDestroys | walkDistance | weaponsAcquired | winPoints | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 9329eb41e215eb | 676b23c24e70d6 | 45b576ab7daa7f | 0 | 0 | 51.46875 | 0 | 0 | 0 | 73 | ... | 1500 | 0 | 0.0 | 0 | 0.0 | 0 | 0 | 588.0 | 1 | 0 |
| 1 | 639bd0dcd7bda8 | 430933124148dd | 42a9a0b906c928 | 0 | 4 | 179.12500 | 0 | 0 | 2 | 11 | ... | 1503 | 2 | 4668.0 | 0 | 0.0 | 0 | 0 | 2017.0 | 6 | 0 |
| 2 | 63d5c8ef8dfe91 | 0b45f5db20ba99 | 87e7e4477a048e | 1 | 0 | 23.40625 | 0 | 0 | 4 | 49 | ... | 1565 | 0 | 0.0 | 0 | 0.0 | 0 | 0 | 788.0 | 4 | 0 |
| 3 | cf5b81422591d1 | b7497dbdc77f4a | 1b9a94f1af67f1 | 0 | 0 | 65.50000 | 0 | 0 | 0 | 54 | ... | 1465 | 0 | 0.0 | 0 | 0.0 | 0 | 0 | 1812.0 | 3 | 0 |
| 4 | ee6a295187ba21 | 6604ce20a1d230 | 40754a93016066 | 0 | 4 | 330.25000 | 1 | 2 | 1 | 7 | ... | 1480 | 1 | 0.0 | 0 | 0.0 | 0 | 0 | 2964.0 | 4 | 0 |
5 rows × 28 columns
pred = lr.predict(X_last_test)
df_sub['winPlacePerc'] = pred
df_sub.to_csv("submission.csv", index=False)