๐Ÿ’ก WIDA/DACON ๋ถ„๋ฅ˜-ํšŒ๊ท€

[DACON/๊น€์„ธ์—ฐ] ์ฒœ์ฒด ์œ ํ˜• ๋ถ„๋ฅ˜ ๋Œ€ํšŒ๋ฅผ ์œ„ํ•œ ๋„๋ฉ”์ธ ์ง€์‹ ์•Œ์•„๋ณด๊ธฐ

์•Œ ์ˆ˜ ์—†๋Š” ์‚ฌ์šฉ์ž 2023. 3. 16. 21:53

๋จธ์‹  ๋Ÿฌ๋‹์˜ ์ข…๋ฅ˜ ๋ฐ ํŠน์ง•

๋จธ์‹  ๋Ÿฌ๋‹์˜ ์ข…๋ฅ˜ ๋ฐ ํŠน์ง•.pdf
0.25MB

 

1)    ์ง€๋„ํ•™์Šต (supervised learning)

: ์ •๋‹ต๊ณผ ๊ฐ€์ด๋“œ๊ฐ€ ์žˆ์Œ  

: ๊ณผ๊ฑฐ์˜ ๋ฐ์ดํ„ฐ๋ฅผ ํ†ตํ•ด ๋ฐฐ์›€

: ์ข…์†๋ณ€์ˆ˜์™€ ๋…๋ฆฝ๋ณ€์ˆ˜๊ฐ€ ์žˆ์Œ

-๋ถ„๋ฅ˜ (classification)

-ํšŒ๊ท€ (regression)

 

2)    ๋น„์ง€๋„ํ•™์Šต (supervised learning)

: ์ •๋‹ต๊ณผ ๊ฐ€์ด๋“œ๊ฐ€ ์—†์Œ

-๊ตฐ์ง‘ํ™” (clustering)

-๋ณ€ํ™˜ (transform)

-์—ฐ๊ด€ (association)

 

3)    ๊ฐ•ํ™”ํ•™์Šต (reinforce learning)

 

 

 

 

#๋ถ„๋ฅ˜: ์˜ˆ์ธกํ•˜๊ณ  ์‹ถ์€ ์ข…์†๋ณ€์ˆ˜๊ฐ€ ์ด๋ฆ„,๋ฌธ์ž์ผ ๋•Œ

 

<์•Œ๊ณ ๋ฆฌ์ฆ˜>

Decision TreeClassifier

KNeighborsClassifier

LogisiticRegression

SVC

RandomForestClassfier

XGBClassifier

 

<ํ‰๊ฐ€๋ฐฉ๋ฒ•>

accuracy_score

recall_score

precision_score

classification_report

confusion_matrix

 

 

 

#ํšŒ๊ท€: ์˜ˆ์ธกํ•˜๊ณ  ์‹ถ์€ ์ข…์†๋ณ€์ˆ˜๊ฐ€ ์ˆซ์ž์ผ ๋•Œ

 

:ํ•™์Šต๋ฐ์ดํ„ฐ๋ฅผ ์ด์šฉํ•˜์—ฌ ์—ฐ์†์ ์ธ ์ˆซ์ž(๊ฐ’)์„ ์˜ˆ์ธกํ•œ๋‹ค.

 

<์•Œ๊ณ ๋ฆฌ์ฆ˜>

linerRegression

KneighborsRegressor

Decision TressRegressor

SVR

RandomFores+Regressor

XGBRegressor

 

<ํ‰๊ฐ€ ๋ฐฉ๋ฒ•>

mean_absolute_error

mean_squared_error

root mean_squared_error

mean_absolute_percentage_error

rZ_score

 

 

#์ฃผ์š” ํ‚ค์›Œ๋“œ์˜ ์˜๋ฏธ

 

์ฒœ์ฒด ์œ ํ˜• ๋ถ„๋ฅ˜ ๋Œ€ํšŒ 1๋“ฑํŒ€ ์ฒ˜์Œํ•ด๋ด์š” ์ฝ”๋“œ ์„ค๋ช… ์ž๋ฃŒ - DACON

type = Source type (์ฒœ์ฒด์˜ ๋ถ„๋ฅ˜)

fiberID = ๊ด€์ธก์— ์‚ฌ์šฉ๋œ ๊ด‘์„ฌ์œ ์˜ ๊ตฌ๋ถ„์ž

psgMag = Point spread function magnitudes, ๋จผ ์ฒœ์ฒด๋ฅผ ํ•œ ์ ์œผ๋กœ ๊ฐ€์ •ํ•˜์—ฌ ์ธก์ •ํ•œ ๋น›

fiberMag = Fiber magnitudes,
3
์ธ์น˜ ์ง€๋ฆ„์˜ ๊ด‘์„ฌ์œ ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๊ด‘์ŠคํŽ™ํŠธ๋Ÿผ์„ ์ธก์ •, ๊ด‘์„ฌ์œ ๋ฅผ ํ†ต๊ณผํ•˜๋Š” ๋น›์˜ ๋ฐ๊ธฐ

petroMag = Petrosian Magnitudes
์€ํ•˜์ฒ˜๋Ÿผ ๋šœ๋ ทํ•œ ํ‘œ๋ฉด์ด ์—†๋Š” ์ฒœ์ฒด์—์„œ๋Š” ๋น›์˜ ๋ฐ๊ธฐ ์ธก์ • ์–ด๋ ค์›€
์ฒœ์ฒด์˜ ์œ„์น˜,๊ฑฐ๋ฆฌ์— ์ƒ๊ด€์—†์ด ๋น›์˜ ๋ฐ๊ธฐ๋ฅผ ๋น„๊ตํ•˜๊ธฐ ์œ„ํ•œ ์ˆ˜์น˜

modelMag = Model magnitude
์ฒœ์ฒด์ค‘์‹ฌ์œผ๋กœ ํŠน์ •๊ฑฐ๋ฆฌ์˜ ๋ฐ๊ธฐ

magnitude = [์ฒœ๋ฌธ] (๋ณ„์˜ ๊ด‘๋„์— ์˜ํ•œ) ๋“ฑ๊ธ‰

serendipity = ๋œป๋ฐ–์˜ ๋ฐœ๊ฒฌ

 

#ํ•ญ์„ฑ์˜ ์ข…๋ฅ˜

 

ํ•ญ์„ฑ์˜ ์ข…๋ฅ˜ | ํ•ญ์„ฑ | ์ฒœ์ฒด | ์ฒœ๋ฌธํ•™์Šต๊ด€ | ์ฒœ๋ฌธ์šฐ์ฃผ์ง€์‹์ •๋ณด (kasi.re.kr)

๋จธ์‹  ๋Ÿฌ๋‹์˜ ์ข…๋ฅ˜ ๋ฐ ํŠน์ง•