Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
IWLS call 17 July (11 AM 18 July NZ)
Search
Nicolas Fauchereau
July 17, 2014
Science
0
52
IWLS call 17 July (11 AM 18 July NZ)
some thoughts on the collaborative paper on ML approaches to seasonal MSLA forecasting
Nicolas Fauchereau
July 17, 2014
Tweet
Share
More Decks by Nicolas Fauchereau
See All by Nicolas Fauchereau
ICU_189
nicolasf
0
73
ICU_188
nicolasf
0
98
ICU_187
nicolasf
0
67
ICU_186
nicolasf
0
76
Seminar MJO Hamilton
nicolasf
0
51
ICU_185
nicolasf
0
51
ICU_184
nicolasf
1
84
ICU_183
nicolasf
0
92
ICU_182_NDJ_2016
nicolasf
0
87
Other Decks in Science
See All in Science
インフラだけではない MLOps の話 @事例でわかるMLOps 機械学習の成果をスケールさせる処方箋 発売記念
icoxfog417
PRO
2
670
Mechanistic Interpretability の紹介
sohtakahashi
0
510
ICRA2024 速報
rpc
3
5.8k
統計学入門講座 第1回スライド
techmathproject
0
200
重複排除・高速バックアップ・ランサムウェア対策 三拍子そろったExaGrid × Veeam連携セミナー
climbteam
0
170
MoveItを使った産業用ロボット向け動作作成方法の紹介 / Introduction to creating motion for industrial robots using MoveIt
ry0_ka
0
260
最適化超入門
tkm2261
14
3.4k
大規模言語モデルの開発
chokkan
PRO
85
42k
観察研究における因果推論
nearme_tech
PRO
1
130
生成AI による論文執筆サポートの手引き(ワークショップ) / A guide to supporting dissertation writing with generative AI (workshop)
ks91
PRO
0
380
20分で分かる Human-in-the-Loop 機械学習におけるアノテーションとヒューマンコンピューターインタラクションの真髄
hurutoriya
5
2.7k
Causal discovery based on non-Gaussianity and nonlinearity
sshimizu2006
0
220
Featured
See All Featured
The Art of Programming - Codeland 2020
erikaheidi
53
13k
We Have a Design System, Now What?
morganepeng
51
7.4k
A better future with KSS
kneath
238
17k
Automating Front-end Workflow
addyosmani
1367
200k
Agile that works and the tools we love
rasmusluckow
328
21k
Imperfection Machines: The Place of Print at Facebook
scottboms
267
13k
10 Git Anti Patterns You Should be Aware of
lemiorhan
PRO
656
59k
Mobile First: as difficult as doing things right
swwweet
222
9.2k
How To Stay Up To Date on Web Technology
chriscoyier
790
250k
Building Better People: How to give real-time feedback that sticks.
wjessup
366
19k
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
11
900
Bash Introduction
62gerente
610
210k
Transcript
IWLS call 17 July Nicolas Fauchereau Sco6
Stephens
Agenda • Opera>onal forecas>ng and ensembles (Rashed)
• Paper collabora>on (Nico / Sco6 [NIWA])
Paper collabora>on • Suggested )tle: Machine Learning approaches
to the predic1on of seasonal Mean Sea Level Anomalies in the Pacific • To be submi0ed to: ? • Authors: Nicolas Fauchereau, Sco> Stephens, Judith Wells, Rashed Chowdhury, John Marra, William Sweet, Doug Ramsay, ???
Paper structure • Introduc)on – Extreme sea level
– Societal benefits of opera>onal extreme sea level risk calendar – Review of exis>ng ini>a>ves – interest of (sta>s>cal) ML predic>on: Open-‐source, lightweight – … – Possible approaches: • EVT • Signal Decomposi)on (trend + )de + MSLA + high-‐frequency), and forecast individual components • Data and methods – Data sources – Data processing (predictors / predictands) – ML algorithms – Model evalua>on (cross-‐valida>on and metrics) • Results – Regression • OLS • MARS • NN – Classifica>ons • LDA • SVM • RF • Conclusions
• Introduc>on: – Everyone, John leading
• Data and Methods: – Data (predictand): sources, QC, decomposi>on: Judith, Sco>, Rashed – Data (predictand): discre>za>on for classifica>on: Nico – Data (SST): sources, decomposi>on (EOF, ICA): Nico – Methods: ML algorithms, cross-‐valida>on, metrics: Nico, Sco> (NN), Judith, Rashed (LDA) • Results – Regression: • OLS: Nico, Judith • MARS: Nico • NN: Sco6 – Classifica>on • LDA: Rashed, Nico • SVM: Nico • RF: Nico Who does what ?
• Google docs ? • GIT ? •
Other ? How to collaborate ?