Slide 1
Slide 1 text
ϚϚͷҰาΛࢧ͑Δ
σʔλੳίϯϖʹ͓͍ͯ
ಛྔཧʹർฐ͍ͯ͠Δશਓྨʹ͍͍͑ͨ
ʙֶशɾਪύΠϓϥΠϯΛఴ͑ͯʙ
$POOFIJUP*ODᖒর
$POOFIJUP.BSDIÉWPMʙػցֶशɾσʔλੳࢢʙ
Slide 2
Slide 2 text
͜Μʹͪʂ
ϚϚͷҰาΛࢧ͑Δ
Slide 3
Slide 3 text
͍͖ͳΓͰ͕͢
ϚϚͷҰาΛࢧ͑Δ
Slide 4
Slide 4 text
σʔλੳίϯϖʢ,BHHMF
4*(/"5&ͳͲʣ
ฉ͍ͨ͜ͱ͋Δਓʙ!
ϚϚͷҰาΛࢧ͑Δ
Slide 5
Slide 5 text
σʔλੳίϯϖʹࢀՃͨ͜͠ͱ͋Δਓʙ!
ϚϚͷҰาΛࢧ͑Δ
Slide 6
Slide 6 text
ಛྔͷཧͬͯͲ͏ͯ͠·͔͢ʁ
ʢςʔϒϧσʔλʹ͓͍ͯʣ
ϚϚͷҰาΛࢧ͑Δ
Slide 7
Slide 7 text
Α͋͘Δύλʔϯʢ࣮ମݧʣ
ϚϚͷҰาΛࢧ͑Δ
Slide 8
Slide 8 text
Α͋͘ΔύλʔϯʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
w ಛྔ࡞Δ
DPM <Z
"
#
$
%>ˠ<Z
"
#
$
%
&
'>
# e.g)
train['A'] = train['A'].fillna(0)
train['B'] = np.log1p(train['B'])
train['E'] = train['A'] + train['B']
df_group = train.groupby('D')['E'].mean()
train['F'] = train['D'].map(df_group)
<>
ɾ
ɾ
ɾ
Slide 9
Slide 9 text
Α͋͘ΔύλʔϯʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
w ͏ಛྔͷΧϥϜ͚ͩࢦఆ͢Δ
# e.g)
feat_col = ['A', 'C', 'D', 'E', 'F', 'J']
x_train = train[feat_col]
y_train = train['y']
# e.g)
clf.fit(x_train, y_train)
w ֶशͤ͞Δ
<>
<>
ɾ
ɾ
ɾ
Slide 10
Slide 10 text
Α͋͘ΔύλʔϯʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
w ͏ಛྔͷΧϥϜ͚ͩࢦఆ͢Δ
# e.g)
feat_col = ['A', 'C', 'D', 'E', 'F', 'J']
x_train = train[feat_col]
y_train = train['y']
# e.g)
clf.fit(x_train, y_train)
w ֶशͤ͞Δ
<>
<>
ɾ
ɾ
ɾ
͋Ε
b'`ͬͯͲΜͳಛྔ͚ͩͬʁ
Slide 11
Slide 11 text
Α͋͘ΔύλʔϯʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
# e.g)
train['A'] = train['A'].fillna(0)
train['B'] = np.log1p(train['B'])
train['E'] = train['A'] + train['B']
df_group = train.groupby('D')['E'].mean()
train['F'] = train['D'].map(df_group)
<>
# e.g)
feat_col = ['A', 'C', 'D', 'E', 'F', 'J']
x_train = train[feat_col]
y_train = train['y']
# e.g)
clf.fit(x_train, y_train)
<>
<>
Slide 12
Slide 12 text
Α͋͘ΔύλʔϯʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
# e.g)
train['A'] = train['A'].fillna(0)
train['B'] = np.log1p(train['B'])
train['E'] = train['A'] + train['B']
df_group = train.groupby('D')['E'].mean()
train['F'] = train['D'].map(df_group)
<>
# e.g)
feat_col = ['A', 'C', 'D', 'E', 'F', 'J']
x_train = train[feat_col]
y_train = train['y']
# e.g)
clf.fit(x_train, y_train)
<>
<>
ݟ͚ͭͨʂ
ʢOPUFCPPLͷ্ͷํʣ
Slide 13
Slide 13 text
Α͋͘ΔύλʔϯʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
# e.g)
train['A'] = train['A'].fillna(0)
train['B'] = np.log1p(train['B'])
train['E'] = train['A'] + train['B']
df_group = train.groupby('D')['E'].mean()
train['F'] = train['D'].map(df_group)
<>
# e.g)
feat_col = ['A', 'C', 'D', 'E', 'F', 'J']
x_train = train[feat_col]
y_train = train['y']
# e.g)
clf.fit(x_train, y_train)
<>
<>
ݟ͚ͭͨʂ
ʢOPUFCPPLͷ্ͷํʣ
ಛྔ͕গͳ͍߹·ͩϚγ͕ͩɺ
ଟ͘ͳͬͯ͘ΔͱͲΜͳܭࢉͰٻΊͨ
ಛྔ͔ͩͬͨΛ͍͍ͪͪߟ͑Δʢ୳͢ʣ
ͷ݁ߏେมͩ͠ɺ͕͔͔࣌ؒΔ
Slide 14
Slide 14 text
Α͋͘Δύλʔϯͦͷʢ࣮ମݧʣ
ϚϚͷҰาΛࢧ͑Δ
Slide 15
Slide 15 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
Αͬ͠Ό͊ʂΊͬͪΌྑ͍είΞͰͨͥʙʙʙ
͜ͷOPUFCPPLΛ%VQMJDBUFͯ͠ɺͬͱྑ͍Ϟσϧ࡞ͬͪΌ͏ͧʂ
Slide 16
Slide 16 text
ҰํɺOPUFCPPLͷதʜ
ϚϚͷҰาΛࢧ͑Δ
Slide 17
Slide 17 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
<> import numpy as np
import pandas as pd
OPUFCPPLͷத
ɾ
ɾ
ɾ
<> submission.to_csv('submission.csv', index=False)
Slide 18
Slide 18 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
<> import numpy as np
import pandas as pd
OPUFCPPLͷத
ɾ
ɾ
ɾ
ɾ
ɾ
ɾ
<> submission.to_csv('submission.csv', index=False)
Slide 19
Slide 19 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
<> import numpy as np
import pandas as pd
OPUFCPPLͷத
ɾ
ɾ
ɾ
ɾ
ɾ
ɾ
<> submission.to_csv('submission.csv', index=False)
Slide 20
Slide 20 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
<> import numpy as np
import pandas as pd
OPUFCPPLͷத
ɾ
ɾ
ɾ
ɾ
ɾ
ɾ
<> submission.to_csv('submission.csv', index=False)
ಉ͡ܭࢉΛԿΒͳ͍ͱ͍͚ͳ͍
ʴ
ºʢ ʣˠແବ
Slide 21
Slide 21 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
ॏͳΔ%VQMJDBUFʹΑΓɺOPUFCPPLࠈʹؕΔՄೳੑʜ
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
ʜʜʜʜʜ
Slide 22
Slide 22 text
Α͋͘ΔύλʔϯͦͷʢJQZOCʣ
ϚϚͷҰาΛࢧ͑Δ
ॏͳΔ%VQMJDBUFʹΑΓɺOPUFCPPLࠈʹؕΔՄೳੑʜ
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
ʜʜʜʜʜ
Slide 23
Slide 23 text
ϚϚͷҰาΛࢧ͑Δ
ࠓ͢͜ͱ
Slide 24
Slide 24 text
ϚϚͷҰาΛࢧ͑Δ
σʔλੳίϯϖʹ͓͍ͯ
ಛྔཧʹർฐ͍ͯ͠Δશਓྨʹ͍͍͑ͨ
ʙֶशɾਪύΠϓϥΠϯΛఴ͑ͯʙ
Slide 25
Slide 25 text
ΞδΣϯμ
ࣗݾհ
ಛྔཧʹ͍ͭͯ
‣ ྻ͝ͱʹQJDLMFϑΝΠϧͰಛྔΛཧ
‣ ಛྔੜ࣌ɺಉ࣌ʹϝϞϑΝΠϧੜ
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
‣ ίϚϯυҰൃͰֶशˠ4VCNJUϑΝΠϧ࡞·ͰΛ࣮ߦ
‣ ֶशʹ༻ͨ͠ಛྔϞσϧύϥϝʔλMPHͱҰॹʹอଘ
‣ TIBQΛ༻͍ͯಛྔͷߩݙΛՄࢹԽ͠ɺ࣍ճֶश࣌ͷצॴΛݟ͚ͭΔ
ϚϚͷҰาΛࢧ͑Δ
Slide 26
Slide 26 text
ΞδΣϯμ
ࣗݾհ
ಛྔཧʹ͍ͭͯ
‣ ྻ͝ͱʹQJDLMFϑΝΠϧͰಛྔΛཧ
‣ ಛྔੜ࣌ɺಉ࣌ʹϝϞϑΝΠϧੜ
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
‣ ίϚϯυҰൃͰֶशˠ4VCNJUϑΝΠϧ࡞·ͰΛ࣮ߦ
‣ ֶशʹ༻ͨ͠ಛྔϞσϧύϥϝʔλMPHͱҰॹʹอଘ
‣ TIBQΛ༻͍ͯಛྔͷߩݙΛՄࢹԽ͠ɺ࣍ճֶश࣌ͷצॴΛݟ͚ͭΔ
ϚϚͷҰาΛࢧ͑Δ
ݰਓͷܙΛ͓आΓͨ͠Β
ΊͬͪΌΑ͔ͬͨ
ʢ˞ʣ
͍ͬͯ͏Λ͠·͢
ʢ˞ʣ͋͘·Ͱओ؍Ͱ͢
Slide 27
Slide 27 text
ࣗݾհ
ϚϚͷҰาΛࢧ͑Δ
Slide 28
Slide 28 text
ࣗݾհ
ϚϚͷҰาΛࢧ͑Δ
໊લɿᖒরʢ/P[BXB5BLBOPCVʣ
ॴଐɿίωώτגࣜձࣾ
ɹɹɿ͔ͨͺ͍!UBLBQZ
w ʙίωώτʹ.-ΤϯδχΞͱͯ͠+0*/
w ػցֶशʢ/-1ɺਪનγεςϜʣΛϝΠϯʹΓͭͭΠϯϑϥʢ"84ʣษڧத
w ,BHHMFͨ͠ΓɺϒϩάʢIUUQTXXXUBLBQZXPSLʣॻ͍ͨΓɺٿͨ͠Γɺ
ϥʔϝϯ৯ͨΓ͍ͯ͠·͢
Slide 29
Slide 29 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
˞ԼههࣄΛࢀߟʹ͍͖ͤͯͨͩ͞·ͨ͠ɻ
ɾ,BHHMFͰ͑Δ'FBUIFSܗࣜΛར༻ͨ͠ಛྔཧ๏
IUUQTBNBMPHIBUFCMPKQFOUSZLBHHMFGFBUVSFNBOBHFNFOU
‣ ྻ͝ͱʹQJDLMFϑΝΠϧͰಛྔΛཧ
‣ ಛྔੜ࣌ɺಉ࣌ʹϝϞϑΝΠϧੜ
Slide 30
Slide 30 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
˞ԼههࣄΛࢀߟʹ͍͖ͤͯͨͩ͞·ͨ͠ɻ
ɾ,BHHMFͰ͑Δ'FBUIFSܗࣜΛར༻ͨ͠ಛྔཧ๏
IUUQTBNBMPHIBUFCMPKQFOUSZLBHHMFGFBUVSFNBOBHFNFOU
‣ ྻ͝ͱʹQJDLMFϑΝΠϧͰಛྔΛཧ
‣ ಛྔੜ࣌ɺಉ࣌ʹϝϞϑΝΠϧੜ
࠷ॳʹΠϝʔδΛڞ༗͠·͢
Slide 31
Slide 31 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
lྻ͝ͱzʹಛྔΛQJDLMFϑΝΠϧͰཧ͢Δ
Slide 32
Slide 32 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
4VSWJWFE 1DMBTT 4FY "HF &NCBSLFE
NBMF 4
NBMF $
GFNBMF $
NBMF $
GFNBMF $
GFNBMF 4
NBMF 4
lྻ͝ͱzʹಛྔΛQJDLMFϑΝΠϧͰཧ͢Δ
Slide 33
Slide 33 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
4VSWJWFE 1DMBTT 4FY "HF &NCBSLFE
NBMF 4
NBMF $
GFNBMF $
NBMF $
GFNBMF $
GFNBMF 4
NBMF 4
TVSWJWFE@USBJOQLM
QDMBTT@USBJOQLM
QDMBTT@UFTUQLM
TFY@USBJOQLM
TFY@UFTUQLM
BHF@USBJOQLM
BHF@UFTUQLM
FNCBSLFE@USBJOQLM
FNCBSLFE@UFTUQLM
lྻ͝ͱzʹಛྔΛQJDLMFϑΝΠϧͰཧ͢Δ
Slide 34
Slide 34 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ಛྔੜ࣌ɺಉ࣌ʹಛྔϝϞΛ࡞͢Δ
Slide 35
Slide 35 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ಛྔੜ࣌ɺಉ࣌ʹಛྔϝϞΛ࡞͢Δ
Slide 36
Slide 36 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ಛྔੜ࣌ɺಉ࣌ʹಛྔϝϞΛ࡞͢Δ
݁ߏେมͦ͏ɾɾɾ
Slide 37
Slide 37 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ಛྔੜ࣌ɺಉ࣌ʹಛྔϝϞΛ࡞͢Δ
ͰɺQZUIPOεΫϦϓτΛ࣮ͭߦ͢Δ͚ͩɻ
Slide 38
Slide 38 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
IPHFQZΛίϚϯυϥΠϯ͔Β࣮ߦ͢Δ͚ͩ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
class Sex(Feature):
def create_features(self):
self.train['Sex'] = train['Sex']
self.test['Sex'] = test['Sex']
create_memo('Sex','ੑผ')
class Age(Feature):
def create_features(self):
self.train['Age'] = train['Age']
self.test['Age'] = test['Age']
create_memo('Age','ྸ')
class Age_mis_val_median(Feature):
def create_features(self):
self.train['Age_mis_val_median'] = train['Age'].fillna(train['Age'].median())
self.test['Age_mis_val_median'] = test['Age'].fillna(test['Age'].median())
create_memo('Age_mis_val_median','ྸͷܽଛΛதԝͰิͨ͠ͷ')
Slide 39
Slide 39 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
class Sex(Feature):
def create_features(self):
self.train['Sex'] = train['Sex']
self.test['Sex'] = test['Sex']
create_memo('Sex','ੑผ')
class Age(Feature):
def create_features(self):
self.train['Age'] = train['Age']
self.test['Age'] = test['Age']
create_memo('Age','ྸ')
class Age_mis_val_median(Feature):
def create_features(self):
self.train['Age_mis_val_median'] = train['Age'].fillna(train['Age'].median())
self.test['Age_mis_val_median'] = test['Age'].fillna(test['Age'].median())
create_memo('Age_mis_val_median','ྸͷܽଛΛதԝͰิͨ͠ͷ')
IPHFQZΛίϚϯυϥΠϯ͔Β࣮ߦ͢Δ͚ͩ
Slide 40
Slide 40 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
class Sex(Feature):
def create_features(self):
self.train['Sex'] = train['Sex']
self.test['Sex'] = test['Sex']
create_memo('Sex','ੑผ')
class Age(Feature):
def create_features(self):
self.train['Age'] = train['Age']
self.test['Age'] = test['Age']
create_memo('Age','ྸ')
class Age_mis_val_median(Feature):
def create_features(self):
self.train['Age_mis_val_median'] = train['Age'].fillna(train['Age'].median())
self.test['Age_mis_val_median'] = test['Age'].fillna(test['Age'].median())
create_memo('Age_mis_val_median','ྸͷܽଛΛதԝͰิͨ͠ͷ')
֤ಛྔ
IPHFQZΛίϚϯυϥΠϯ͔Β࣮ߦ͢Δ͚ͩ
Slide 41
Slide 41 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
class Sex(Feature):
def create_features(self):
self.train['Sex'] = train['Sex']
self.test['Sex'] = test['Sex']
create_memo('Sex','ੑผ')
class Age(Feature):
def create_features(self):
self.train['Age'] = train['Age']
self.test['Age'] = test['Age']
create_memo('Age','ྸ')
class Age_mis_val_median(Feature):
def create_features(self):
self.train['Age_mis_val_median'] = train['Age'].fillna(train['Age'].median())
self.test['Age_mis_val_median'] = test['Age'].fillna(test['Age'].median())
create_memo('Age_mis_val_median','ྸͷܽଛΛதԝͰิͨ͠ͷ')
ಛྔϝϞϑΝΠϧ
IPHFQZΛίϚϯυϥΠϯ͔Β࣮ߦ͢Δ͚ͩ
Slide 42
Slide 42 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
DSFBUF@NFNPͷॲཧ֓ཁ
Slide 43
Slide 43 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
DSFBUF@NFNPͷॲཧ֓ཁ
# ಛྔϝϞcsvϑΝΠϧ࡞
def create_memo(col_name, desc):
file_path = Feature.dir + '/_features_memo.csv'
if not os.path.isfile(file_path):
with open(file_path,"w"):pass
with open(file_path, 'r+') as f:
lines = f.readlines()
lines = [line.strip() for line in lines]
# ॻ͖ࠐ͏ͱ͍ͯ͠Δಛྔ͕͢Ͱʹॻ͖ࠐ·Ε͍ͯͳ͍͔νΣοΫ
col = [line for line in lines if line.split(',')[0] == col_name]
if len(col) != 0:return
writer = csv.writer(f)
writer.writerow([col_name, desc])
Slide 44
Slide 44 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
DSFBUF@NFNPͷॲཧ֓ཁ
# ಛྔϝϞcsvϑΝΠϧ࡞
def create_memo(col_name, desc):
file_path = Feature.dir + '/_features_memo.csv'
if not os.path.isfile(file_path):
with open(file_path,"w"):pass
with open(file_path, 'r+') as f:
lines = f.readlines()
lines = [line.strip() for line in lines]
# ॻ͖ࠐ͏ͱ͍ͯ͠Δಛྔ͕͢Ͱʹॻ͖ࠐ·Ε͍ͯͳ͍͔νΣοΫ
col = [line for line in lines if line.split(',')[0] == col_name]
if len(col) != 0:return
writer = csv.writer(f)
writer.writerow([col_name, desc])
Slide 45
Slide 45 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Pclass(Feature):
def create_features(self):
self.train['Pclass'] = train['Pclass']
self.test['Pclass'] = test['Pclass']
create_memo('Pclass','νέοτͷΫϥεɻ1st, 2nd, 3rdͷ3छྨ')
DSFBUF@NFNPͷॲཧ֓ཁ
# ಛྔϝϞcsvϑΝΠϧ࡞
def create_memo(col_name, desc):
file_path = Feature.dir + '/_features_memo.csv'
if not os.path.isfile(file_path):
with open(file_path,"w"):pass
with open(file_path, 'r+') as f:
lines = f.readlines()
lines = [line.strip() for line in lines]
# ॻ͖ࠐ͏ͱ͍ͯ͠Δಛྔ͕͢Ͱʹॻ͖ࠐ·Ε͍ͯͳ͍͔νΣοΫ
col = [line for line in lines if line.split(',')[0] == col_name]
if len(col) != 0:return
writer = csv.writer(f)
writer.writerow([col_name, desc])
$47ܗࣜͰอଘ͓ͯ͘͠ͱ(JUIVC͔Βࢀর͍͢͠
ʢͪΖΜɺ&YDFM/VNCFSTͱ͍ͬͨΞϓϦέʔγϣϯ͔ΒͰ៉ྷʹݟ͑Δʣ
Slide 46
Slide 46 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
৽͍͠ಛྔΛ࡞͢Δ߹
Slide 47
Slide 47 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Family_Size(Feature):
def create_features(self):
self.train['Family_Size'] = train['Parch'] + train['SibSp']
self.test['Family_Size'] = test['Parch'] + test['SibSp']
create_memo('Family_Size','Ոͷ૯')
IPHFQZʹ৽͍͠ಛྔੜॲཧΛهड़
৽͍͠ಛྔΛ࡞͢Δ߹
Slide 48
Slide 48 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Family_Size(Feature):
def create_features(self):
self.train['Family_Size'] = train['Parch'] + train['SibSp']
self.test['Family_Size'] = test['Parch'] + test['SibSp']
create_memo('Family_Size','Ոͷ૯')
QZUIPOIPHFQZ
৽͍͠ಛྔΛ࡞͢Δ߹
IPHFQZʹ৽͍͠ಛྔੜॲཧΛهड़
Slide 49
Slide 49 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
class Family_Size(Feature):
def create_features(self):
self.train['Family_Size'] = train['Parch'] + train['SibSp']
self.test['Family_Size'] = test['Parch'] + test['SibSp']
create_memo('Family_Size','Ոͷ૯')
QZUIPOIPHFQZ
৽͍͠ಛྔͷΈੜ
৽͍͠ಛྔΛ࡞͢Δ߹
IPHFQZʹ৽͍͠ಛྔੜॲཧΛهड़
Slide 50
Slide 50 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
σʔλΛಡΈࠐΉࡍɺಛྔΛࢦఆͯ͠ϩʔυ͢Δ͚ͩ
# ಛྔͷࢦఆ
features = [
"age_mis_val_median",
"family__size",
"cabin",
"fare_mis_val_median"
]
df = [pd.read_pickle(FEATURE_DIR_NAME + f’{f}_train.pkl') for f in features]
df = pd.concat(df, axis=1)
Slide 51
Slide 51 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
σʔλΛಡΈࠐΉࡍɺಛྔΛࢦఆͯ͠ϩʔυ͢Δ͚ͩ
# ಛྔͷࢦఆ
features = [
"age_mis_val_median",
"family__size",
"cabin",
"fare_mis_val_median"
]
df = [pd.read_pickle(FEATURE_DIR_NAME + f’{f}_train.pkl') for f in features]
df = pd.concat(df, axis=1)
Կ͕خ͔͔ͬͨ͠
Slide 52
Slide 52 text
ಛྔཧʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
w ͭͷεΫϦϓτϑΝΠϧʹಛྔੜΛ·ͱΊΔ͜ͱͰɺಉ͡ܭࢉΛෳ
ճ࣮ߦ͢Δ͜ͱΛආ͚ɺ࣌ؒΛ༗ޮ׆༻Ͱ͖Δɻ
ɹˠಛྔͷ࠶ݱੑ୲อɻ
w ಛྔͷϝϞΛಉ࣌ʹੜ͢Δ͜ͱͰʮ͜ͷಛྔͳΜ͚ͩͬʁʯͱ಄Λ
ΘͣʹࡁΜͩɻ
w ಛྔΛྻ͝ͱʹཧ͢Δ͜ͱͰऔΓճָ͕͠ʹͳΔɻ
ɹˠQJDLMFϑΝΠϧͩͱอଘಡΈࠐΈ͍ʂ
ɹˠಛྔ͕େʹͳΔ߹ɺ͋Δఔͷ୯ҐͰ·ͱΊͯཧ͢Δํ͕ྑ͍͔ɻ
Slide 53
Slide 53 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
˞Լهॻ੶ʹܝࡌ͞Ε͍ͯΔύΠϓϥΠϯΛࢀߟʹ͠·ͨ͠ɻ
ɾ,BHHMFͰউͭσʔλੳͷٕज़
IUUQTHJIZPKQCPPL
‣ ίϚϯυҰൃͰֶशˠ4VCNJUϑΝΠϧ࡞·ͰΛ࣮ߦ
‣ ֶशʹ༻ͨ͠ಛྔϞσϧύϥϝʔλMPHͱҰॹʹอଘ
‣ TIBQΛ༻͍ͯಛྔͷߩݙΛՄࢹԽ͠ɺ࣍ճֶश࣌ͷצॴΛݟ͚ͭΔ
Slide 54
Slide 54 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
SVOQZΛ࣮ߦ͢Δ͜ͱͰɺֶशɾਪɾ4VCNJUϑΝΠϧΛ࡞
# ಛྔͷࢦఆ
features = [
"age_mis_val_median",
"family__size",
"cabin",
"fare_mis_val_median"
]
run_name = 'lgb_1102'
# ༻͢ΔಛྔϦετͷอଘ
with open(LOG_DIR_NAME + run_name + "_features.txt", 'wt') as f:
for ele in features:
f.write(ele+'\n')
params_lgb = {
'boosting_type': 'gbdt',
'objective': 'binary',
'early_stopping_rounds': 20,
'verbose': 10,
'random_state': 99,
'num_round': 100
}
# ༻͢Δύϥϝʔλͷอଘ
with open(LOG_DIR_NAME + run_name + "_param.txt", 'wt') as f:
for key,value in sorted(params_lgb.items()):
f.write(f'{key}:{value}\n')
runner = Runner(run_name, ModelLGB, features, params_lgb, n_fold, name_prefix)
runner.run_train_cv() # ֶश
runner.run_predict_cv() # ਪ
Submission.create_submission(run_name) # submit࡞
Slide 55
Slide 55 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
SVOQZΛ࣮ߦ͢Δ͜ͱͰɺֶशɾਪɾ4VCNJUϑΝΠϧΛ࡞
# ಛྔͷࢦఆ
features = [
"age_mis_val_median",
"family__size",
"cabin",
"fare_mis_val_median"
]
run_name = 'lgb_1102'
# ༻͢ΔಛྔϦετͷอଘ
with open(LOG_DIR_NAME + run_name + "_features.txt", 'wt') as f:
for ele in features:
f.write(ele+'\n')
params_lgb = {
'boosting_type': 'gbdt',
'objective': 'binary',
'early_stopping_rounds': 20,
'verbose': 10,
'random_state': 99,
'num_round': 100
}
# ༻͢Δύϥϝʔλͷอଘ
with open(LOG_DIR_NAME + run_name + "_param.txt", 'wt') as f:
for key,value in sorted(params_lgb.items()):
f.write(f'{key}:{value}\n')
runner = Runner(run_name, ModelLGB, features, params_lgb, n_fold, name_prefix)
runner.run_train_cv() # ֶश
runner.run_predict_cv() # ਪ
Submission.create_submission(run_name) # submit࡞
͜ͷSVO@OBNFΛQSFpYͱͯ͠ɺϑΝΠϧϞσϧΛอଘͯ͘͠ΕΔɻ
ྫ
w ༻ͨ͠ಛྔϦετ
w ༻ͨ͠ϋΠύʔύϥϝʔλ
w GPMEຖͷϞσϧ
w ਪ݁Ռ
w TVCNJUϑΝΠϧ
w TIBQͷܭࢉ݁ՌΠϝʔδϑΝΠϧͳͲ
Slide 56
Slide 56 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
SVOQZΛ࣮ߦ͢Δ͜ͱͰɺֶशɾਪɾ4VCNJUϑΝΠϧΛ࡞
# ಛྔͷࢦఆ
features = [
"age_mis_val_median",
"family__size",
"cabin",
"fare_mis_val_median"
]
run_name = 'lgb_1102'
# ༻͢ΔಛྔϦετͷอଘ
with open(LOG_DIR_NAME + run_name + "_features.txt", 'wt') as f:
for ele in features:
f.write(ele+'\n')
params_lgb = {
'boosting_type': 'gbdt',
'objective': 'binary',
'early_stopping_rounds': 20,
'verbose': 10,
'random_state': 99,
'num_round': 100
}
# ༻͢Δύϥϝʔλͷอଘ
with open(LOG_DIR_NAME + run_name + "_param.txt", 'wt') as f:
for key,value in sorted(params_lgb.items()):
f.write(f'{key}:{value}\n')
runner = Runner(run_name, ModelLGB, features, params_lgb, n_fold, name_prefix)
runner.run_train_cv() # ֶश
runner.run_predict_cv() # ਪ
Submission.create_submission(run_name) # submit࡞
ੜ͞ΕΔϑΝΠϧྫ
Slide 57
Slide 57 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ੜ͞ΕΔϑΝΠϧͷྫʢϑΥϧμద͚͍ٓͯ·͢ʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@QSFEQLMʢUFTUσʔλͰͷਪ݁Ռʣ
w MHC@@@TVCNJTTJPODTWʢਪ݁ՌΛLBHHMFʹఏग़Ͱ͖ΔDTWʹมͨ͠ͷʣ
w MHC@@@GFBUVSFTUYUʢࠓճͷֶशʹ༻ͨ͠ಛྔϦετʣ
w MHC@@@QBSBNUYUʢࠓճͷֶशʹ༻ͨ͠ϋΠύʔύϥϝʔλʣ
w MHC@@@TIBQQOHʢTIBQͰܭࢉͨ͠ՄࢹԽΠϝʔδʣ
w HFOFSBMMPHʢܭࢉϩάϑΝΠϧʣ
w SFTVMUMPHʢϞσϧͷείΞ͚͕ͩهࡌ͞ΕͨϩάϑΝΠϧʣ
Slide 58
Slide 58 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ੜ͞ΕΔϑΝΠϧͷྫʢϑΥϧμద͚͍ٓͯ·͢ʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@QSFEQLMʢUFTUσʔλͰͷਪ݁Ռʣ
w MHC@@@TVCNJTTJPODTWʢਪ݁ՌΛLBHHMFʹఏग़Ͱ͖ΔDTWʹมͨ͠ͷʣ
w MHC@@@GFBUVSFTUYUʢࠓճͷֶशʹ༻ͨ͠ಛྔϦετʣ
w MHC@@@QBSBNUYUʢࠓճͷֶशʹ༻ͨ͠ϋΠύʔύϥϝʔλʣ
w MHC@@@TIBQQOHʢTIBQͰܭࢉͨ͠ՄࢹԽΠϝʔδʣ
w HFOFSBMMPHʢܭࢉϩάϑΝΠϧʣ
w SFTVMUMPHʢϞσϧͷείΞ͚͕ͩهࡌ͞ΕͨϩάϑΝΠϧʣ
Slide 59
Slide 59 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ੜ͞ΕΔϑΝΠϧͷྫʢϑΥϧμద͚͍ٓͯ·͢ʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@QSFEQLMʢUFTUσʔλͰͷਪ݁Ռʣ
w MHC@@@TVCNJTTJPODTWʢਪ݁ՌΛLBHHMFʹఏग़Ͱ͖ΔDTWʹมͨ͠ͷʣ
w MHC@@@GFBUVSFTUYUʢࠓճͷֶशʹ༻ͨ͠ಛྔϦετʣ
w MHC@@@QBSBNUYUʢࠓճͷֶशʹ༻ͨ͠ϋΠύʔύϥϝʔλʣ
w MHC@@@TIBQQOHʢTIBQͰܭࢉͨ͠ՄࢹԽΠϝʔδʣ
w HFOFSBMMPHʢܭࢉϩάϑΝΠϧʣ
w SFTVMUMPHʢϞσϧͷείΞ͚͕ͩهࡌ͞ΕͨϩάϑΝΠϧʣ
Slide 60
Slide 60 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ੜ͞ΕΔϑΝΠϧͷྫʢϑΥϧμద͚͍ٓͯ·͢ʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@QSFEQLMʢUFTUσʔλͰͷਪ݁Ռʣ
w MHC@@@TVCNJTTJPODTWʢਪ݁ՌΛLBHHMFʹఏग़Ͱ͖ΔDTWʹมͨ͠ͷʣ
w MHC@@@GFBUVSFTUYUʢࠓճͷֶशʹ༻ͨ͠ಛྔϦετʣ
w MHC@@@QBSBNUYUʢࠓճͷֶशʹ༻ͨ͠ϋΠύʔύϥϝʔλʣ
w MHC@@@TIBQQOHʢTIBQͰܭࢉͨ͠ՄࢹԽΠϝʔδʣ
w HFOFSBMMPHʢܭࢉϩάϑΝΠϧʣ
w SFTVMUMPHʢϞσϧͷείΞ͚͕ͩهࡌ͞ΕͨϩάϑΝΠϧʣ
Slide 61
Slide 61 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ੜ͞ΕΔϑΝΠϧͷྫʢϑΥϧμద͚͍ٓͯ·͢ʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@QSFEQLMʢUFTUσʔλͰͷਪ݁Ռʣ
w MHC@@@TVCNJTTJPODTWʢਪ݁ՌΛLBHHMFʹఏग़Ͱ͖ΔDTWʹมͨ͠ͷʣ
w MHC@@@GFBUVSFTUYUʢࠓճͷֶशʹ༻ͨ͠ಛྔϦετʣ
w MHC@@@QBSBNUYUʢࠓճͷֶशʹ༻ͨ͠ϋΠύʔύϥϝʔλʣ
w MHC@@@TIBQQOHʢTIBQͰܭࢉͨ͠ՄࢹԽΠϝʔδʣ
w HFOFSBMMPHʢܭࢉϩάϑΝΠϧʣ
w SFTVMUMPHʢϞσϧͷείΞ͚͕ͩهࡌ͞ΕͨϩάϑΝΠϧʣ
Slide 62
Slide 62 text
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
ϚϚͷҰาΛࢧ͑Δ
ੜ͞ΕΔϑΝΠϧͷྫʢϑΥϧμద͚͍ٓͯ·͢ʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@GPMENPEFMʢGPMEͰ࡞͞ΕͨϞσϧʣ
w MHC@@QSFEQLMʢUFTUσʔλͰͷਪ݁Ռʣ
w MHC@@@TVCNJTTJPODTWʢਪ݁ՌΛLBHHMFʹఏग़Ͱ͖ΔDTWʹมͨ͠ͷʣ
w MHC@@@GFBUVSFTUYUʢࠓճͷֶशʹ༻ͨ͠ಛྔϦετʣ
w MHC@@@QBSBNUYUʢࠓճͷֶशʹ༻ͨ͠ϋΠύʔύϥϝʔλʣ
w MHC@@@TIBQQOHʢTIBQͰܭࢉͨ͠ՄࢹԽΠϝʔδʣ
w HFOFSBMMPHʢܭࢉϩάϑΝΠϧʣ
w SFTVMUMPHʢϞσϧͷείΞ͚͕ͩهࡌ͞ΕͨϩάϑΝΠϧʣ
Կ͕خ͔͔ͬͨ͠
Slide 63
Slide 63 text
ϚϚͷҰาΛࢧ͑Δ
w ʮ͜ͷಛྔʯͱʮ͜ͷύϥϝʔλʯΛֶͬͯशͤͨ͞Ϟσϧ
ʹؔͯ͠ɺʮ֤λεΫʹཁͨ࣌ؒ͠ʯͱʮ֤GPMEʴ࠷ऴతͳεί
ΞʯΛҙࣝ͠ͳͯ͘ཧͰ͖ΔΑ͏ʹɻ
w TIBQͷܭࢉ݁ՌGFBUVSFJNQPSUBODFΛग़ྗ͓ͯ͘͜͠ͱ
Ͱɺ࣍ͷֶश࣌ͷצॴ͕௫ΊΔΑ͏ʹɻ
ֶशɾਪύΠϓϥΠϯʹ͍ͭͯ
Slide 64
Slide 64 text
·ͱΊ
ϚϚͷҰาΛࢧ͑Δ
Slide 65
Slide 65 text
·ͱΊ
ϚϚͷҰาΛࢧ͑Δ
w ಛྔཧ͍͍ͧʂ
ͭͷεΫϦϓτϑΝΠϧʹಛྔੜΛ·ͱΊΔ͜ͱͰɺಉ͡ܭࢉΛෳճ࣮ߦ͢Δ͜ͱ
ΛճආͰ͖Δʂ
ಛྔͷϝϞΛಉ࣌ʹੜ͢Δ͜ͱͰʮ͜ͷಛྔͳΜ͚ͩͬʁʯͱ಄Λ͏ճ͕ݮ
Δʂ
ಛྔΛྻ͝ͱʹཧ͢Δ͜ͱͰऔΓճָ͕͠ʹͳͬͨʂʢ͕ɺಛྔ͕େͳ߹͋
Δఔͷ·ͱ·ΓͰཧͨ͠ํ͕ྑ͍͔ʣ
w ύΠϓϥΠϯ͍͍ͧʂ
ύΠϓϥΠϯΛߏங͢Δ͜ͱͰɺߴͳ1%$"Λ࣮ݱʂ
ֶशʹ༻ͨ͠ಛྔͱύϥϝʔλΛཧ͢Δ͜ͱͰɺ࠶ݱੑ୲อ͞Ε৺ཧత҆શੑ
Slide 66
Slide 66 text
ϚϚͷҰาΛࢧ͑Δ
͝ਗ਼ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠