FISH TALES - hea- · 2/8/2006  · If you teach a man to fish he will eat for a lifetime. FISH...

Post on 15-Jul-2020

0 views 0 download

Transcript of FISH TALES - hea- · 2/8/2006  · If you teach a man to fish he will eat for a lifetime. FISH...

If you give a man a fish he willeat for a day. If you teach aman to fish he will eat for alifetime.

FISH TALES

If you give a man Cash he willanalyze for a day. If you teach aman Poisson, he will analyze fora lifetime.

POISSON TALES

It is all about Probabilities!

•The Poisson Likelihood

•Probability Calculus – Bayes’ Theorem

•Priors – the gamma distribution

•Source intensity and background marginalization

•Hardness Ratios (BEHR)

Deriving the Poisson Likelihood

Deriving the Poisson Likelihood

Deriving the Poisson Likelihood

Deriving the Poisson Likelihood

10

Ν,τ→∞

Deriving the Poisson Likelihood

10

Ν,τ→∞

Deriving the Poisson Likelihood

Probability Calculus

A or B :: p(A+B) = p(A) + p(B) - p(AB)

A and B :: p(AB) = p(A|B) p(B) = p(B|A) p(A)

Bayes’ Theorem :: p(B|A) = p(A|B) p(B) / p(A)

p(Model |Data) = p(Model) p(Data|Model) / p(Data)

posteriordistribution

priordistribution

likelihood normalization

marginalization :: p(a|D) = p(ab|D) db

Priors

Priors

• Incorporate known information

• Forced acknowledgement of bias

• Non-informative priors

– flat

– range

– least informative (Jeffrey’s)

– gamma

Priors

the gamma distribution

Priors

the gamma distribution

Panning for Gold:Source and Background

Panning for Gold:Source and Background

Panning for Gold:Source and Background

Hardness Ratios

Simple, robust, intuitive summaryProxy for spectral fittingUseful for large samples

Most needed for low counts

Hardness Ratios

Simple, robust, intuitive summaryProxy for spectral fittingUseful for large samples

Most needed for low counts

BEHRhttp://hea-www.harvard.edu/AstroStat/BEHR/