Apresentao do PowerPoint

1
Principal Component Analysis (PCA) of the ultra
high energy extensive air showers longitudinal
profile
Fernando Catalani, Jun Takahashi e José Augusto
Chinellato DRCC IFGW UNICAMP
2

• Our knowledge of the mass composition of the
  cosmic rays of highest energy (above 0.1EeV) is
  deficient.
• Measurement of the mass of this cosmic rays is
  necessary to understand their origin and
  propagation.
• Top-Down models predict mainly photons and
  protons while Botton-Up models propose protons
  and heavy nuclei.

3

• The detection of shower maximum depth via
  fluorescence is one method to study mass
  composition.
• The study of Longitudinal Profile of extensive
  air showers with the fluorescent light technique
  was implemented by the Fly's eye experiment.
• Recently the Fluorescence Detector (FD) of Pierre
  Auger has started to collect data.
• A large amount of data will be produced and new
  and accurate data analysis procedures are
  welcome.

4
PCA Principal Component Analysis

• The pca is a very useful tool widely used.
• The basic idea is to project the raw data onto a
  few dominant eigenvectors with large variance in
  order to reduce the dimension, noise of the
  original data.
• The new basis provides a simple description and
  can reveal underlying characteristics of the
  original data set.

5
Parameters Nmax number of particles at the
maximum. Xmax- depth of the shower maximum. s -
width of profile. Asymmetry of the
profile 4000 CORSIKA Showers
Nmax
s
Xmax
6
Distribution of Xmax
Energy 1.1018 eV
7
Distribution of Xmax and parameter f4 (PCA)
MF(?Fe- ?p)/sqrt(?fe2 ?p2)
Energy 1.1018 eV
8
Iron and Proton separation in 1018 eV and 1017 eV
p3 PCA applied over proton f4 PCA applied
over iron
9
Distribution of Xmax and parameter f4 (PCA)
Energy 1.1017 eV
10
(No Transcript)
11
Xmax for different energies
12
f4 parameter for different energies
13
Merit Factor X Energy
14
Conclusions

• PCA over FD parameters (Nmax, Xmax, ?, and
  assymetry) brings better results concerning mass
  composition than usual Xmax analysis.
• Further improvements will take into account other
  primaries than iron and proton.
• PCA over surface detector variables.

15
(No Transcript)
16
(No Transcript)