Add Hlt2RD_LbToPpKmPipPimMuMu_MVA line
Add Hlt2RD_LbToPpKmPipPimMuMu_MVA line. Used for searching for rare decay \Lambda_{b}\to p K^-\pi^+\pi^-\mu^+\mu^-, optimized using MVA method
Bandwidth and rate
Tested with scripts in rd-rta-guide/testing-utils
Only this line:
| file size | bandwidth | rate | |
|---|---|---|---|
| No MVA | 0.22 MB | 2.98 MB/s | 108.44 Hz |
| With MVA | 0.0 MB | 0.0 MB/s. | 0.0 Hz |
All RD turbo lines:
| file size | bandwidth | |
|---|---|---|
| rd-devel-june25 | 30 MB | 406.6413 MB/s |
| This MR | 30 MB | 406.6413 MB/s |
Got same total bandwidth. The rate comparison is here: compare_rate.log
Selection
Much looser cut compare to Hlt2RD_LbToPpKmPipPimMuMu, especially pions cuts
Background suppressed by a BDT
All cuts are summarized in following tables:
| Particles | Variable | Value |
|---|---|---|
| All final states | isLong | true |
| Particles | Variable | Value |
|---|---|---|
| Muon | isMuon | true |
| p_{\mathrm{T}} | >250\mathrm{MeV}/c | |
| p | >3000\mathrm{MeV}/c | |
| \chi^2_{\mathrm{IP}} | >4 | |
| \mathrm{PID}_{\mu} | >2 |
| Particles | Variable | Value |
|---|---|---|
| DiMuon | Mass | [200, 5400]\mathrm{MeV}/c^2 |
| p_{\mathrm{T}} | >0\mathrm{MeV}/c | |
| \chi^2_{\mathrm{DOCA}} | <20 | |
| \chi^2_{\mathrm{vtx}}/\mathrm{ndof} | <10 | |
| \chi^2_{\mathrm{FD}} | >20 |
| Particles | Variable | Value |
|---|---|---|
| Pion | hasRich | true |
| p_{\mathrm{T}} | >0\mathrm{MeV}/c | |
| \chi^2_{\mathrm{IP}} | >4 | |
| \mathrm{PID}_{K} | <-1 |
| Particles | Variable | Value |
|---|---|---|
| DiPion | Mass | [0, 4000]\mathrm{MeV}/c^2 |
| p_{\mathrm{T}} | >0\mathrm{MeV}/c | |
| \chi^2_{\mathrm{DOCA}} | <20 | |
| \chi^2_{\mathrm{vtx}}/\mathrm{ndof} | <9 |
| Particles | Variable | Value |
|---|---|---|
| Proton | hasRich | true |
| p_{\mathrm{T}} | >300\mathrm{MeV}/c | |
| p | >9\mathrm{GeV}/c | |
| \chi^2_{\mathrm{IP}} | >4 | |
| \mathrm{PID}_{p} | >0 | |
| \mathrm{PID}_{p}-\mathrm{PID}_{K} | >-2 |
| Particles | Variable | Value |
|---|---|---|
| Kaon | hasRich | true |
| p_{\mathrm{T}} | >250\mathrm{MeV}/c | |
| p | >3.0\mathrm{GeV}/c | |
| \chi^2_{\mathrm{IP}} | >4 | |
| \mathrm{PID}_{K} | >0 |
| Particles | Variable | Value |
|---|---|---|
| pK^-\pi^+\pi^- | Mass | [0, 5.5]\mathrm{GeV}/c^2 |
| \Sigma p_{\mathrm{T}} | >1.5\mathrm{GeV}/c | |
| \Sigma \chi^{2}_{\mathrm{IP}} | >20.0 | |
| \mathrm{max}(\chi^{2}_{\mathrm{DOCA}}) | < 20 | |
| \chi^2_{\mathrm{vtx}}/\mathrm{ndof} | < 16 |
| Particles | Variable | Value |
|---|---|---|
| \Lambda_b | Mass | [4800, 6900]\mathrm{MeV}/c^2 |
| p_{\mathrm{T}} | >2500\mathrm{MeV}/c | |
| \chi^2_{\mathrm{vtx}} | <25 | |
| \mathrm{DIRA} | >0.995 | |
| \chi^2_{\mathrm{IP}} | <25 | |
| \chi^2_{\mathrm{FD}} | >40 | |
| \chi^2_{\mathrm{DOCA}} | >25 |
BDT
BDT trained with:
- Bkg proxy: Hlt2RD_LbToPpKmPipPimMuMu sideband <5100 MeV, >6100 MeV
- Sig proxy: 2016,2017,2018 MC
Train variables:
- p_{\mathrm{T}} of all final states except pions
- Logarithm of \chi^2_{\mathrm{IP}} of all final states and \Lambda_b
- Sum of p_{\mathrm{T}} of hadronic final states
- Logarithm of sum of \chi^2_{\mathrm{IP}} of all hadronic final states
Working point: -0.1
Weight file: lhcb-datapkg/ParamFiles!154 (merged)
BDT response:
Edited by Qi Shi
