Skip to content
This repository has been archived by the owner on Sep 19, 2022. It is now read-only.

be-favorite/Research_archive

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

45 Commits
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation

Research Archive

๋…ผ๋ฌธ๊ณผ ์ฑ…, ์›น์‚ฌ์ดํŠธ ๋“ฑ์„ ํ†ตํ•ด ๊ณต๋ถ€ํ•˜๊ณ  ์—ฐ๊ตฌํ•œ ๊ฒƒ๋“ค์„ ์•„์นด์ด๋ธŒํ•ฉ๋‹ˆ๋‹ค.

๊ทธ๋ฆฌ๊ณ , ๊ทธ ์™ธ Data Science๋ฅผ ํ•  ๋•Œ ์•Œ์•„๋‘๋ฉด ์ข‹๋‹ค๊ณ  ์ƒ๊ฐ๋˜๋Š” ๊ฒƒ๋“ค๋„ ์•„์นด์ด๋ธŒํ•ฉ๋‹ˆ๋‹ค.

์ฐธ๊ณ  ๋ฌธํ—Œ๊ณผ ์Šคํ„ฐ๋”” ๋…ธํŠธ, ๊ทธ๋ฆฌ๊ณ  ๊ฐ€๋Šฅํ•˜๋‹ค๋ฉด ์žฌํ˜„๊ฐ€๋Šฅํ•œ ์ฝ”๋“œ ๋˜๋Š” ์žฌํ˜„๊ฐ€๋Šฅํ•œ ๊ฐ„๋žตํ•œ ํŠœํ† ๋ฆฌ์–ผ์„ ํ•จ๊ป˜ ์ œ๊ณตํ•˜๊ณ ์ž ํ•ฉ๋‹ˆ๋‹ค.

๊ณต๋ถ€ํ•˜๊ณ  ์—ฐ๊ตฌํ–ˆ๋˜ ํฐ ์ฃผ์ œ๋“ค์ž…๋‹ˆ๋‹ค:

Time Series

1 ์ถ”๋ก  ๋ชจ๋ธ๋ง ยท Regression

Spurious regression

  • ๋‚˜์ข…ํ™”. R ์‘์šฉ ์‹œ๊ณ„์—ด๋ถ„์„. ์ž์œ ์•„์นด๋ฐ๋ฏธ. 2020.
  • ์—ฌ๋Ÿฌ ์‹œ๊ณ„์—ด๋กœ ํšŒ๊ท€๋ฅผ ์ˆ˜ํ–‰ํ•  ๋•Œ, ๊ผญ ์ฃผ์˜ํ•ด์•ผ ํ•  ์•Œ์•„๋‘์–ด์•ผํ•  ์‚ฌํ•ญ
  • ๐Ÿ”— ์Šคํ„ฐ๋”” ๋…ธํŠธ
  • ๐Ÿ”— R ํŠœํ† ๋ฆฌ์–ผ: CCF ๋ถ„์„์˜ ํ—ˆ๊ตฌ์  ์ƒ๊ด€ ํ™•์ธ ๊ณผ์ • ์ฐธ๊ณ 

Regression with ARIMA errors

Distributed lag model

Distributed lag non-linear model

2 ์˜ˆ์ธก๋ชจ๋ธ๋ง ยท Forecasting

Exponential Smoothing

ARIMA model

Prophet

Hierarchical Time Series Forecasting

3 Other techniques

Intervention analysis (Interrupted Time Series)

Dynamic Time Warping (DTW)

  • Berndt, Donald J., and James Clifford. โ€œUsing Dynamic Time Warping to Find Patterns in Time Series.โ€ In Proceedings of the 3rd International Conference on Knowledge Discovery and Data Mining, 359โ€“70. AAAIWSโ€™94. Seattle, WA: AAAI Press, 1994.
  • ์„ ํ–‰ ๋˜๋Š” ํ›„ํ–‰ํ•˜๋Š” ์‹œ๊ณ„์—ด, ์‹œ์ฐจ๊ฐ€ ์กด์žฌํ•˜๋‚˜ ์œ ์‚ฌํ•œ ํŒจํ„ด์ด ์กด์žฌํ•˜๋Š” ๋‘ ์‹œ๊ณ„์—ด์„ ์žก์•„๋‚ผ ์ˆ˜ ์žˆ๊ฒŒ๋” ํ•ด์ฃผ๋Š” ๋น„์œ ์‚ฌ์„ฑ ์ธก๋„(๊ฑฐ๋ฆฌ ์ธก๋„) ์•Œ๊ณ ๋ฆฌ์ฆ˜
  • DTW distance๋ฅผ ์ด์šฉํ•ด ๊ณ„์ธต์  ๊ตฐ์ง‘ ๋ถ„์„ ์ˆ˜ํ–‰ ๊ฐ€๋Šฅ
  • ๐Ÿ”— ์Šคํ„ฐ๋”” ๋…ธํŠธ
  • ๐Ÿ”— R ํŠœํ† ๋ฆฌ์–ผ

Discrete Wavelet Transform (DWT)

  • Graps, Amara. โ€œAn Introduction to Wavelets.โ€ IEEE Comp. Sci. Engi. 2 (February 1, 1995): 50โ€“61. https://doi.org/10.1109/99.388960.
  • Li, Daoyuan, Tegawendรฉ F. Bissyandรฉ, Jacques Klein, and Y. L. Traon. โ€œTime Series Classification with Discrete Wavelet Transformed Data: Insights from an Empirical Study.โ€ In SEKE, 2016. https://doi.org/10.18293/SEKE2016-067.
  • ์‹œ๊ณ„์—ด๋“ค์„ ๋ฐ์ดํ„ฐ์˜ ์—ด๋กœ ๋‚˜์—ดํ•˜์—ฌ classification์„ ์ˆ˜ํ–‰ํ•  ๋•Œ, ํšจ๊ณผ์ ์ธ ์ฐจ์› ๊ฐ์†Œ ๋ฐฉ๋ฒ•
  • ์ผ์ข…์˜ ์‹œ๊ณ„์—ด Feature engineering ๊ธฐ๋ฒ•์— ํ•ด๋‹น
  • ๐Ÿ”— ์Šคํ„ฐ๋”” ๋…ธํŠธ
  • ๐Ÿ”— R ํŠœํ† ๋ฆฌ์–ผ

Machine Learning and Statistical Learning

Prerequisite

Ensemble methods

Logistic regression

Generalized Linear Model (GLM) and Generalized Additive Model (GAM)

Deep Learning

Prerequisites

High-Dimensional Data Analysis

  • Breheny, Patrick. High-Dimensional Data Analysis. The University of Iowa, 2016. https://myweb.uiowa.edu/pbreheny/7600/s16/index.html.
    • R ์†Œ์Šค์ฝ”๋“œ ๋ฐ ์˜ˆ์ œ Dataset ์ œ๊ณต
  • ์ผ๋ฐ˜์ ์ธ ๊ธฐ๊ณ„ํ•™์Šต ๊ธฐ๋ฐ˜์˜ ์˜ˆ์ธก ๋ชจ๋ธ๋ง์œผ๋กœ ์ ‘๊ทผํ•˜๊ธฐ ์–ด๋ ค์šด n -> p ๋˜๋Š” n < p ์ธ ์ž๋ฃŒ์˜ ์˜ˆ์ธก ๋ชจ๋ธ๋ง์— ๊ด€ํ•œ ๋ฐฉ๋ฒ•๋ก (์—ฌ๊ธฐ์„œ n์€ ๊ด€์ธก์น˜์˜ ์ˆ˜, p๋Š” ์˜ˆ์ธก๋ณ€์ˆ˜์˜ ์ˆ˜)
  • ๊ผญ ๊ณ ์ฐจ์› ์ž๋ฃŒ๊ฐ€ ์•„๋‹Œ, ํšŒ๊ท€๋ชจํ˜•์˜ ์˜ˆ์ธก ์„ฑ๋Šฅ์„ ๋†’์ด๊ธฐ ์œ„ํ•ด์„œ๋„ ์‚ฌ์šฉ๋˜๋Š” ๋ฐฉ๋ฒ•๋ก ๋“ค์— ํ•ด๋‹น
  • ํ†ต๊ณ„์  ๊ฐ€์„ค๊ฒ€์ • ๊ด€์ ์—์„œ ๊ฐ€์„ค ๊ฒ€์ •์‹œ ๋ฐœ์ƒํ•˜๋Š” ๊ณ ์ฐจ์› ๋ฌธ์ œ์— ๊ด€ํ•œ ์†”๋ฃจ์…˜ ๋˜ํ•œ ์ œ๊ณตํ•จ

1 ๊ณ ์ฐจ์› ์ž๋ฃŒ์— ๊ด€ํ•œ ์˜ˆ์ธก ๋ชจ๋ธ๋ง

Prerequisites

Ridge regression

Lasso regression

Bias reduction of Lasso estimator

Variance reduction of Lasso eistimator

Penalized logistic regression

Penalized robust regression

2 ํ†ต๊ณ„์  ๊ฐ€์„ค๊ฒ€์ • ๊ด€์ ์˜ ๊ณ ์ฐจ์› ๋ฌธ์ œ

Prerequisites

Family-Wise Error Rates (FWER)

False Discovery Rates (FDR)

Statistics

  • ํ†ต๊ณ„ํ•™, ํ†ต๊ณ„์  ๊ฐ€์„ค๊ฒ€์ •๊ณผ ๊ด€๋ จํ•œ ๊ฒƒ๋“ค์„ ์•„์นด์ด๋ธŒ ํ•ฉ๋‹ˆ๋‹ค.

๊ตฌ๊ฐ„์ถ”์ •์˜ ํ•ด์„์— ๋Œ€ํ•œ ๊ณ ์ „์  ๊ด€์ (Frequentist)๊ณผ ๋ฒ ์ด์ง€์•ˆ ๊ด€์ 

๊ฒ€์ •๋ ฅ(power)๊ณผ ๊ฒ€์ •๋ ฅ ํ•จ์ˆ˜์— ๋Œ€ํ•ด

์ž์œ ๋„(Degrees of Freedom)

ํ‘œ์ค€ํŽธ์ฐจ์™€ ํ‘œ์ค€์˜ค์ฐจ

"๋Œ€๋ฆฝ๊ฐ€์„ค์ด ์˜ณ๋‹ค."๋ผ๋Š” ์‹์˜ ์ฃผ์žฅ์„ ์ง€์–‘ํ•ด์•ผํ•˜๋Š” ์ด์œ 

์ค‘์‹ฌ๊ทนํ•œ์ •๋ฆฌ์˜ ์˜๋ฏธ

Fixed effect์™€ random effect

Miscellaneous

๊ฒฐ์ •๋ก ์  SIR ๋ชจํ˜•์„ ์ด์šฉํ•œ ๊ฐ์—ผ๋ณ‘ ์œ ํ–‰ ๋ชจ๋ธ๋ง