Here is a website with a lot of good links.
http://user.pa.msu.edu/linnemann/public/workshop/stat_resources.html
Thursday, June 30, 2011
Wednesday, June 29, 2011
6/29
A lot of stuff today. Andy worked with me almost all day, which was astounding. He also gave me several papers to review and a very good process to go to build up to be doing the error analysis we want to be doing. I better go to catch the bus.
Overall goal now is to get very familar with xspec, get intuitive feel for how changing the different parameters of the models changes the fit, how exposure times change the errors, fit parameters, and other factors. devote completely to xspec, terminal, statistics. Keep on the track he wrote down.
Overall goal now is to get very familar with xspec, get intuitive feel for how changing the different parameters of the models changes the fit, how exposure times change the errors, fit parameters, and other factors. devote completely to xspec, terminal, statistics. Keep on the track he wrote down.
6/30:
Recommended by Andy:
numerical recipes in C++
http://www.nr.com/
reading, section 15.5 non-linear models
I. confidence limits
[cofidence limits] summarizes the probability distribution of errors in parameter estimation.
" The full probability distrubtion is a function defined on the M-dimensional space of parameters a
numerical recipes in C++
http://www.nr.com/
reading, section 15.5 non-linear models
I. confidence limits
[cofidence limits] summarizes the probability distribution of errors in parameter estimation.
" The full probability distrubtion is a function defined on the M-dimensional space of parameters a
Tuesday, June 28, 2011
6/28
links
confidence interval wikipedia
http://xassist.pha.jhu.edu/pipeline4/xmm/0304850901/xmm0304850901/report/fullrun_mos1_xmm0304850901_pi300-10000_pass1_cl_src3.html
-total angular momentum
http://www.chemglobe.org/ptoe/_/14.php
using help in general is very useful for getting help (woaaaaaah)
confidence interval wikipedia
http://xassist.pha.jhu.edu/pipeline4/xmm/0304850901/xmm0304850901/report/fullrun_mos1_xmm0304850901_pi300-10000_pass1_cl_src3.html
-total angular momentum
http://www.chemglobe.org/ptoe/_/14.php
using help in general is very useful for getting help (woaaaaaah)
Monday, June 27, 2011
6/27/2011: First day
Well, a ton of information. Whew, I will have to get good sleep to stay on top of everything.
Recap of day
-Met Dr. Ptak
-Talked about limitations of CCD astronomy, spectral resolution (error goes like 1/sqrt(N)
- N is the number of electrons, which is determined by photon energy/ ionization energy (ionization energy is energy needed to remove one electron from an atom).
- talked about effective area, about how it is similar to optical telescopes
-looked at Chandra website, Education, in order understand how x-ray telescopes work, with their diffration gratings
-talked about the point spread function, which shows how much a point source is spread out from thermal noise, electronic noise
-When looking at spectra, looked actual data for x-ray flux and the model for x-ray counts, using a simple POWER LAW fit
F(E)= N*E^(-a)
-this is related to counts, as F(E)/E= Counts= N*E^(-a)-1= N*E^-(c), where c=a+1
taking log of both sides,
log P(E) is proportional to -c*log(E) + log(n)
SO, it should be linear. However, it is not. (See journal).
GOALS:
model X-Ray sources, analyze errors
Dr. Ptak mentioned another good idea was to develop program to automatically run modelling, given different response functions
-Calorimeters have better spectral resolution than CCD's, CCD's work by ionization of atoms on the detector, usually Si, Ge.
Photon energy is proportional to ionization energy
LOOK UP: Gaussian, Poission distributions!!
review Standard deviation, look this up in data analysis book!
Potential to do: simulate radiactive emission of Cl-38 to model spectra of Cl-38, using random library. I am not very sure how to start doing this.
-Dr. Ptak then brought up some basic optics to show how the focal length of the lense (10m) and the pixel size impacts the angular resolution.
This allows you to figure out the source size on the detector, pi*(angular resolution)^2 (angular resolution is radius?)
Dr. Ptak then installed heasoft. He used many shortcuts in the terminal to make the process quicker.
In order to get HEASOFT to work, he had to add 3 keywords into several files. I believe this was included in the log of programming of today, in the folder documents/6-27
He mentioned that arf essentially is the mirror, while rmf is essentially the detector.
keep reading about what arf, rmf are.
LOOK UP: what files did Dr. Krane use to collect data?
phabs= photon absorption
FOR TOMORROW:
play with terminal
look up cross-section theory
play with x-spec
get card from office secretary on 2nd floor
read about fits, file extensions
read about gaussian, poisson distributions.
review everything we went over today
TO ASK Dr. Ptak
1. What was he doing exactly when he was doing this step process?
XSPEC12>step 2 2.4 2.6 10
C-Statistic Delta PhoIndex
C-Statistic 2
404.28 0.093251 0 2.4
404.39 0.20187 1 2.42
404.54 0.35127 2 2.44
404.73 0.54097 3 2.46
404.96 0.77047 4 2.48
405.23 1.0389 5 2.5
405.54 1.3458 6 2.52
405.88 1.6904 7 2.54
406.26 2.0718 8 2.56
406.68 2.4893 9 2.58
407.13 2.9422 10 2.6
XSPEC12>step 2 2.4 2.8 10
ask about how to add the blank files keywords to the files (the programming wasn't saved).
Here are the links that to some of the pages Andy showed me, just to make sure I have a link to them.
This is a link that will show me the locations of different buildings at Goddard
http://wikimapia.org/#lat=38.993005&lon=-76.853571&z=16&l=0&m=b
This is a link to arf, rmf, and other files for the a source from data taken from NGC 253, this is the SOURCE 3 data.
This link gives the equation for determining cross section absorption
http://xassist.pha.jhu.edu/pipeline4/xmm/0110900101/xmm0110900101/report/fullrun_mos2_xmm0110900101_pi300-10000_pass1_cl_src3.html
Recap of day
-Met Dr. Ptak
-Talked about limitations of CCD astronomy, spectral resolution (error goes like 1/sqrt(N)
- N is the number of electrons, which is determined by photon energy/ ionization energy (ionization energy is energy needed to remove one electron from an atom).
- talked about effective area, about how it is similar to optical telescopes
-looked at Chandra website, Education, in order understand how x-ray telescopes work, with their diffration gratings
-talked about the point spread function, which shows how much a point source is spread out from thermal noise, electronic noise
-When looking at spectra, looked actual data for x-ray flux and the model for x-ray counts, using a simple POWER LAW fit
F(E)= N*E^(-a)
-this is related to counts, as F(E)/E= Counts= N*E^(-a)-1= N*E^-(c), where c=a+1
taking log of both sides,
log P(E) is proportional to -c*log(E) + log(n)
SO, it should be linear. However, it is not. (See journal).
GOALS:
model X-Ray sources, analyze errors
Dr. Ptak mentioned another good idea was to develop program to automatically run modelling, given different response functions
-Calorimeters have better spectral resolution than CCD's, CCD's work by ionization of atoms on the detector, usually Si, Ge.
Photon energy is proportional to ionization energy
LOOK UP: Gaussian, Poission distributions!!
review Standard deviation, look this up in data analysis book!
Potential to do: simulate radiactive emission of Cl-38 to model spectra of Cl-38, using random library. I am not very sure how to start doing this.
-Dr. Ptak then brought up some basic optics to show how the focal length of the lense (10m) and the pixel size impacts the angular resolution.
This allows you to figure out the source size on the detector, pi*(angular resolution)^2 (angular resolution is radius?)
Dr. Ptak then installed heasoft. He used many shortcuts in the terminal to make the process quicker.
In order to get HEASOFT to work, he had to add 3 keywords into several files. I believe this was included in the log of programming of today, in the folder documents/6-27
He mentioned that arf essentially is the mirror, while rmf is essentially the detector.
keep reading about what arf, rmf are.
LOOK UP: what files did Dr. Krane use to collect data?
phabs= photon absorption
FOR TOMORROW:
play with terminal
look up cross-section theory
play with x-spec
get card from office secretary on 2nd floor
read about fits, file extensions
read about gaussian, poisson distributions.
review everything we went over today
TO ASK Dr. Ptak
1. What was he doing exactly when he was doing this step process?
XSPEC12>step 2 2.4 2.6 10
C-Statistic Delta PhoIndex
C-Statistic 2
404.28 0.093251 0 2.4
404.39 0.20187 1 2.42
404.54 0.35127 2 2.44
404.73 0.54097 3 2.46
404.96 0.77047 4 2.48
405.23 1.0389 5 2.5
405.54 1.3458 6 2.52
405.88 1.6904 7 2.54
406.26 2.0718 8 2.56
406.68 2.4893 9 2.58
407.13 2.9422 10 2.6
XSPEC12>step 2 2.4 2.8 10
ask about how to add the blank files keywords to the files (the programming wasn't saved).
Here are the links that to some of the pages Andy showed me, just to make sure I have a link to them.
This is a link that will show me the locations of different buildings at Goddard
http://wikimapia.org/#lat=38.993005&lon=-76.853571&z=16&l=0&m=b
This is a link to arf, rmf, and other files for the a source from data taken from NGC 253, this is the SOURCE 3 data.
This link gives the equation for determining cross section absorption
http://xassist.pha.jhu.edu/pipeline4/xmm/0110900101/xmm0110900101/report/fullrun_mos2_xmm0110900101_pi300-10000_pass1_cl_src3.html
Monday, June 20, 2011
Topics, Concepts to Research
1. Moseley's Work
2. K, L, M Shells in electrons
3. Electron transitions (K-alpha, beta, etc.)
4. J-vector, quantum numbers
5. spin, implications to atomic energy levels
2. K, L, M Shells in electrons
3. Electron transitions (K-alpha, beta, etc.)
4. J-vector, quantum numbers
5. spin, implications to atomic energy levels
Goal of Blog
-Keep log of X-Ray telescope internship
-keep note of thesis ideas I have
-->keep lists of topics/ concepts to research, books/articles to read, and questions I have.
---->To do this, I should take note of it in the journal, and then write it up more neatly on the computer.
I think I should try and keep the following progression for writing about my work:
1. Journal (Rough Draft)
2. Word Document (final draft)
3. Blog (published draft)
To save time, skip the word document step if it is not necessary. I need to practice trying to communicate my ideas, discoveries to others. This will allow others to read what I have been doing, and give good documentation that will help me in writing my thesis and communicating more clearly what I have done to others. I feel like I can have better organization with writing about my progress on-line.
-keep note of thesis ideas I have
-->keep lists of topics/ concepts to research, books/articles to read, and questions I have.
---->To do this, I should take note of it in the journal, and then write it up more neatly on the computer.
I think I should try and keep the following progression for writing about my work:
1. Journal (Rough Draft)
2. Word Document (final draft)
3. Blog (published draft)
To save time, skip the word document step if it is not necessary. I need to practice trying to communicate my ideas, discoveries to others. This will allow others to read what I have been doing, and give good documentation that will help me in writing my thesis and communicating more clearly what I have done to others. I feel like I can have better organization with writing about my progress on-line.
Subscribe to:
Comments (Atom)