Kernel PCA for SNe Kernel PCA for SNe photometric classification - - PowerPoint PPT Presentation

▶
kernel pca for sne kernel pca for sne photometric
SMART_READER_LITE
LIVE PREVIEW

Kernel PCA for SNe Kernel PCA for SNe photometric classification - - PowerPoint PPT Presentation

Dec 05, 2023 15 likes •156 views

Kernel PCA for SNe Kernel PCA for SNe photometric classification photometric classification Emille E. O. Ishida IAG University of Sao Paulo, Brazil In collaboration with Rafael S. De Souza (KASI) astro-ph/1201.6676 IAU General Assembly,

slide-1
SLIDE 1

IAU General Assembly, August/2012, Beijing, China

Kernel PCA for SNe Kernel PCA for SNe photometric classification photometric classification

astro-ph/1201.6676

Emille E. O. Ishida

IAG – University of Sao Paulo, Brazil

In collaboration with Rafael S. De Souza (KASI)

slide-2
SLIDE 2

IAU General Assembly, August/2012, Beijing, China

T ype Ia SN surveys are not able to provide complete spectroscopic follow up

  • 1. The problem
  • 1. The problem

We have too much data!

Ironically...

slide-3
SLIDE 3

IAU General Assembly, August/2012, Beijing, China

IIP

  • 1. The problem
  • 1. The problem
slide-4
SLIDE 4

IAU General Assembly, August/2012, Beijing, China

Which one is a Ia?

  • 1. The problem
  • 1. The problem
slide-5
SLIDE 5

IAU General Assembly, August/2012, Beijing, China

Dimensionality reduction technique Look for directions that maximizes variance

  • 2. Principal Component Analysis (PCA)
  • 2. Principal Component Analysis (PCA)

http://web.media.mit.edu/~tristan/phd/dissertation/figures/PCA.jpg

slide-6
SLIDE 6

IAU General Assembly, August/2012, Beijing, China

  • 2. PCA limitations
  • 2. PCA limitations

Does not care about labels Is not designed to capture non-linear structure

slide-7
SLIDE 7

IAU General Assembly, August/2012, Beijing, China

Sometimes, going to higher dimensions might solve the problem

  • 2. PCA extensions
  • 2. PCA extensions
slide-8
SLIDE 8

IAU General Assembly, August/2012, Beijing, China First natural choice - Gaussian kernel

2.The kernel trick 2.The kernel trick

slide-9
SLIDE 9

IAU General Assembly, August/2012, Beijing, China First natural choice - Gaussian kernel

2.The kernel trick 2.The kernel trick

slide-10
SLIDE 10

IAU General Assembly, August/2012, Beijing, China

Classification Classification

THE Nearest Neighbor (1NN)

  • 3. kPCA applied to SNe classification
  • 3. kPCA applied to SNe classification
slide-11
SLIDE 11

IAU General Assembly, August/2012, Beijing, China

Selection cuts: {-3,+24} in r-band At least 3 obs with SNR>5 in each band

FoM ~0.60 SC ~91%

  • 3. kPCA applied to SNe classification
  • 3. kPCA applied to SNe classification

P

  • s

t

  • S

N P C C s a m p l e A f t e r s e l e c t i

  • n

c u t s

slide-12
SLIDE 12

IAU General Assembly, August/2012, Beijing, China

  • 3. kPCA applied to SNe classification
  • 3. kPCA applied to SNe classification

Results from the Supernova Photometric Classification Challenge (Kessler et al., 2010)

Our results Our results

slide-13
SLIDE 13

IAU General Assembly, August/2012, Beijing, China

  • 3. kPCA applied to SNe classification
  • 3. kPCA applied to SNe classification

Results from the Supernova Photometric Classification Challenge (Kessler et al., 2010)

average purity: 75% average purity: 75%

Same result as winner SNPCC (76%), without using host redshift information

hostZ No hostZ Better result in intermediate redshift Our results Our results

slide-14
SLIDE 14

IAU General Assembly, August/2012, Beijing, China

  • 4. Conclusions
  • 4. Conclusions
  • 1. SNe photometric classification is not a future issue..

it is already here!

  • 3. There is no need of enviromental, redshift or astrophysical

hypothesis

  • 2. kPCA is a powerfull tool, mainly if we are interested

in a high quality purity in intermediate redshifts.

  • 4. Great potential in detecting previously non-observed objects:

Application to PISN search