BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Nuclei Identification by Multiple Energy Losses in Detectors of th
 e PAMELA Spectrometer
DTSTART;VALUE=DATE-TIME:20220715T070000Z
DTEND;VALUE=DATE-TIME:20220715T072000Z
DTSTAMP;VALUE=DATE-TIME:20261017T212808Z
UID:indico-contribution-481@events.sinp.msu.ru
DESCRIPTION:Speakers: Vladislav Alekseev (Yaroslavl State University)\nPAM
 ELA magnetic spectrometer [1] includes a set of detectors\, each of which 
 measures the energy losses of charged particles along its trajectory. In t
 otal\, it is possible to obtain up to 6 values of energy loss in detectors
  of the time-of-flight system\, up to 12 values in the track system\, and 
 up to 43 values in the calorimeter. This information combined with rigidit
 y and velocity measurements can be used for effective identification of pa
 rticles.\nWe propose two algorithms for identification of nuclei with atom
 ic numbers 1 – 26 (from protons up to iron): the first algorithm is base
 d on maximum likelihood estimation method\, and the second is based on mac
 hine learning. For both algorithms we performed Monte Carlo simulation of 
 particles interaction in detectors of PAMELA with Geant4 software [2].\nTh
 e first algorithm consists of the following steps.\n1. Fit the energy loss
  distributions in simulation for each detector depending on particle type 
 and magnetic rigidity $R$ by a 2-parametric family of distributions $f(x\;
  \\mu\, \\sigma)$. \n2. Approximation of the relation between parameters $
 (\\mu\, \\sigma)$ and particle rigidity by some functions $\\hat{\\mu}(x)$
  and $\\hat{\\sigma}(x)$. As a result\, we obtain a set of functions $\\ha
 t{\\mu}_{ij}(x)$ and $\\hat{\\sigma}_{ij}(x)$ for each nucleus number (cha
 rge) $i$ and detector number $j$.\n3. After that\, we can apply an algorit
 hm to the experimental data. For each nucleus type calculate a likelihood 
 function (where $R$ is magnetic rigidity):\n$$L_i(R) = \\sum\\limits_{j = 
 1}{N} \\ln f(R\; \\hat{\\mu}_{ij}(R)\, \\hat{\\sigma}_{ij}(R))$$\nThe resu
 lt of an identification algorithm is a nucleus $n$\, which minimizes $L_i 
 (R)$:\n$$\\hat{n} = \\arg\\min\\limits_{i}⁡L_i(R).$$\nThe second method 
 is based on machine learning using the gradient boosting method\, which is
  currently one of the reference methods of classical learning [3]. The mag
 netic rigidity and energy releases along the trajectory are used as featur
 es.\nWe present the results of each of the above methods separately\, as w
 ell as in comparison with each other.\n1.	Adriani O. et al. Astrophys J (2
 014) 791:93\n2.	S. Agostinelli et al.\, Nucl. Instrum. Meth. A 506\, pp. 2
 50-303\, 2003.\n3.	Friedman\, Jerome. (2000). Greedy Function Approximatio
 n: A Gradient Boosting Machine. The Annals of Statistics. 29.10.1214/aos/1
 013203451\n\nhttps://events.sinp.msu.ru/event/8/contributions/481/
LOCATION: Физический ф-т\, СФА
URL:https://events.sinp.msu.ru/event/8/contributions/481/
END:VEVENT
END:VCALENDAR