• Aucun résultat trouvé

Showing spectral variability in AGN by principal component analysis (PCA)

N/A
N/A
Protected

Academic year: 2021

Partager "Showing spectral variability in AGN by principal component analysis (PCA)"

Copied!
25
0
0

Texte intégral

(1)

Showing spectral variability in AGN

by principal component analysis (PCA)

G. Miniutti and G. Risalitti

Beatriz Agís González

(2)

INTRODUCTION

1.1- What is PCA?

1.2- Example

1.3- Simulations

NGC4051

2.1- Previous analaysis with PCA

2.2- Another observation

2.3- Further analysis: greatest flux

2.4- Furhter analysis: lowest flux

CONCLUSIONS AND FUTURE WORK

(3)

1. INTRODUCTION

LINEAR TRANSFORMATION METHOD

n variables

correlate amog

each other

WHAT IS PCA?

p < n

uncorrelated

variables

INITIAL DATA SET

NEW DATA SET

that yields the directions

that maximizethe variance of the data

ANALYSING SPECTRA :

Set of

spectra

Individually

variant spectral

componets in a model

independent way

PCA

(4)

y3(x)=sin(2x)

1. INTRODUCTION

EXAMPLE

y

2

(x)=sin(x)

y

1

(x) = 0.3 +x/8

y

i

(x) = 0.6

a

1

y

1

(x)+0.2

a

2

y

2

(x)+0.1

a

3

y

3

(x)+0.1

a

4

Parker et al. (2015)

(5)

DIFFICULT INTERPRETATION

1. INTRODUCTION

SIMULATIONS

Number of variable components of the initial spectrum contributing

to the variability in a

model independent way

SIMULATIONS

Baseline Model :

POWER LAW

Varying

photon index

and

normalization

Timesliced spectra binned in energy

(6)

1. INTRODUCTION

SIMULATIONS

POWER LAW

+

SOFT EXCESS

Parker et al. (2015)

Parker et al. (2015)

-PP:

varying

photon

index

varying

normalization

-BB:

constant

-PP:

varying

normalization

-BB:

varying

normalization

-PP:

varying

photon

index

varying

normalization

-BB

: varying

normalization

(stronger BB variations)

-PP:

varying

photon

index

varying

normalization

-BB:

varying

normalization

(stronger PP variations)

(7)

1. INTRODUCTION

SIMULATIONS

RELATIVISTIC REFLECTION COMPONENT

Parker et al. (2015)

Parker et al. (2015)

Parker et al. (2015)

-PP:

varying

photon index

varying

normalization

-RR:

constant

(8)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

14 OBSERVATIONS

~40ks

(9)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

(10)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

POWER LAW

Varying

photon index

and

normalization

CONSTANT REFLECTION COMPONENT

(11)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

POWER LAW

Varying

photon index

and

normalization

&

RELATIVISTIC REFLECTION

CONSTANT REFLECTION COMPONENT

(12)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

POWER LAW

Varying

photon index

and

normalization

&

RELATIVISTIC REFLECTION

CONSTANT REFECTION COMPONENT

(13)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

VARYING POWER LAW

&

RELATIVISTIC REFLECTION

(14)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

VARYING POWER LAW

&

RELATIVISTIC REFLECTION

CONSTANT REFLECTION COMPONENT

PARTIALLY COVERED

POWER LAW

(15)

PREVIOUS PCA ANALYSIS

=

2. NGC 4051

Parker et al. (2015)

VARYING POWER LAW

&

RELATIVISTIC REFLECTION

CONSTANT REFLECTION COMPONENT

PARTIALLY COVERED

POWER LAW

(16)

ANOTHER OBSERVATION: Same PCs

2. NGC 4051

(17)

ANOTHER OBSERVATION: Same PCs

2. NGC 4051

(18)

Further analysis: GREATEST FLUX

(19)

Further analysis: GREATEST FLUX

(20)

Further analysis: GREATEST FLUX

2. NGC 4051

Parker et al. (2015)

Simulation

Data

Power law varying in normalization

+

(21)

Further analysis: LOWEST FLUX

2. NGC 4051

(22)

Further analysis: LOWEST FLUX

2. NGC 4051

(23)

Power law varying in normalization

+

constant relativistic reflection component

+

Soft Excess

Further analysis: LOWEST FLUX

2. NGC 4051

(24)

Further analysis: LOWEST FLUX

2. NGC 4051

Data

Simulation

Vaughan et al. (2011)

Power law varying in normalization

+

constant relativistic reflection component

+

(25)

Check previuos Parker et al (2015)

Corroborate

constant soft excess

Vaughan et al (2011)

EVOLUTION

of the PCs with

FLUX

Check and simulated intermediate flux states

Check this evolution in other sources

CONCLUSIONS

Références

Documents relatifs

Charged particle acceleration in the course of magnetic re- connection is considered now as one of the main mechanisms producing high-energy electrons in the magnetotail current

Next, according to the interpolation theorem, we calculate the barycentric curves of Chinese, mathematics, geography (respectively shown in Fig.1, Fig.2, Fig.3).

The problems as well as the objective of this realized work are divided in three parts: Extractions of the parameters of wavelets from the ultrasonic echo of the detected defect -

The problems as well as the objective of this realized work are divided in three parts: Extractions of the parameters of wavelets from the ultrasonic echo of the detected defect -

when a part is close to a line (machine), it is assigned to the cell which component this machine. Otherwise, it is an exceptional part which can be illustrated, for example,

Moreover, from the number of photons in X-ray bursts produced by −CGs estimated from ground observations, we show that the proportionality between the number of thermal

Nonlinear principal component analysis (NLPCA) is applied on detrended monthly Sea Surface Temperature Anomaly (SSTA) data from the tropical Atlantic Ocean (30°W-20°E, 26°S-22°..

We show how slight variants of the regularization term of [1] can be used successfully to yield a structured and sparse formulation of principal component analysis for which we