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
DAILY PREDICTION OF PARTICULATE MATTERS USING NEURAL NETWORKS AND ENSEMBLING METHODS
Search
Diego Triana
December 20, 2019
Science
0
47
DAILY PREDICTION OF PARTICULATE MATTERS USING NEURAL NETWORKS AND ENSEMBLING METHODS
Final results of the investigation project.
Diego Triana
December 20, 2019
Tweet
Share
More Decks by Diego Triana
See All by Diego Triana
Preliminary project plan
dtrianab
0
85
Other Decks in Science
See All in Science
DEIM2024 チュートリアル ~AWSで生成AIのRAGを使ったチャットボットを作ってみよう~
yamahiro
3
690
SIGDIAL論文読み会: PGTask: Introducing the Task of Profile Generation from Dialogues
kaiyo3
0
100
Mastering Feature Engineering: Mining the Hidden Salary Formula with CakeResume
tlyu0419
0
130
ABEMAの効果検証事例〜効果の異質性を考える〜
s1ok69oo
3
1.6k
名古屋市立大学データサイエンス学部 夏のオープンキャンパス模擬授業20230818
ncu_ds
0
1.3k
はじめての「相関と因果とエビデンス」入門:“動機づけられた推論” に抗うために
takehikoihayashi
14
4.7k
ChatGPT によるプログラミング授業の課題の解答生成の評価
toskamiya
0
270
WeMeet Group - 採用資料
wemeet
0
160
(Forkwell Library #48)『詳解 インシデントレスポンス』で学び倒すブルーチーム技術
scientia
2
980
Machine Learning for Materials (Lecture 8)
aronwalsh
0
330
Machine Learning for Materials (Lecture 9)
aronwalsh
0
120
Endocannabinology 101
drbonci
PRO
0
260
Featured
See All Featured
Put a Button on it: Removing Barriers to Going Fast.
kastner
58
3.1k
WebSockets: Embracing the real-time Web
robhawkes
59
7k
GraphQLとの向き合い方2022年版
quramy
33
12k
Docker and Python
trallard
35
2.7k
What the flash - Photography Introduction
edds
64
11k
Save Time (by Creating Custom Rails Generators)
garrettdimon
PRO
0
14
Building a Scalable Design System with Sketch
lauravandoore
457
32k
The Brand Is Dead. Long Live the Brand.
mthomps
49
29k
The Pragmatic Product Professional
lauravandoore
26
5.8k
ParisWeb 2013: Learning to Love: Crash Course in Emotional UX Design
dotmariusz
104
6.6k
GraphQLの誤解/rethinking-graphql
sonatard
56
9.3k
How to name files
jennybc
65
93k
Transcript
DAILY PREDICTION OF PARTICULATE MATTERS USING NEURAL NETWORKS AND ENSEMBLING
METHODS By Diego Triana Supervisor: prof. dr hab. inż. Stanisław Osowski 1
PROBLEM DESCRIPTION Measurements of PM in a given place Raw
Data Model 1 MLP Model 2 RBF Model 3 SVM Training Testing FORECAST Can we improve the forecast performance by combining the models? Performance 2
METHODOLOGY • PM2.5 Chinese • PM10 Polish • Data Analysis
▪ Statistical Analysis ▪ Processing input variables • Generation of Diagnostic Features • Feature Selection • PCA • Stepwise • Design of Predictors ▪ MLP ❖ Levenberg-Marquardt back- propagation algorithm ▪ RBF ❖ Levenberg-Marquardt algorithm ▪ SVM ❖ Bayesian optimization (C , σ and ε) ▪ RF ❖ LSBoost • Measure of Performance • Comparison of Results • Ensemble Methods ▪ Dynamic Integration ▪ Bagging ▪ Adaboost • Optimization of Ensembles ▪ Number of Predictors ▪ Loss Function 3
DATASETS (RAW DATA) 4
5
6
DIAGNOSTIC FEATURES 7
8
Daily Chinese Dataset 9
Daily Polish Dataset 10
11
12
PERFORMANCE OF SINGLE PREDICTORS PM2.5 Chinese Data PM10 Polish Data
13
REPEATABILITY OF SINGLE PREDICTORS PM10 (Poland) PM2.5 (China) 14
BAGGING • Bags of 80% from whole training dataset •
150 Predictors 15
ADABOOST Slopes: PM10 |PL | MLP | γ =3.5 PM10
|PL | RF | γ =2.0 PM2.5 |CH| MLP | γ =4.5 PM2.5 |CH| RF | γ =2.5 T=150 Predictors 16
MAE (AVERAGE 20 TIMES RUNS) 17
MAPE (AVERAGE 20 TIMES RUNS) 18
RMSD (AVERAGE 20 TIMES RUNS) 19
Q&A • THANK YOU 20 Instytut Elektrotechniki Teoretycznej i Systemów
Informacyjno-Pomiarowych Zakład Elektrotechniki Teoretycznej i Informatyki Stosowanej