]>
Commit | Line | Data |
---|---|---|
1ee39b3a | 1 | /************************************************************************* |
9ec5bce7 | 2 | * Copyright(c) 1998-1999, ALICE Experiment at CERN, All rights reserved. * |
3 | * * | |
4 | * Author: The ALICE Off-line Project. * | |
5 | * Contributors are mentioned in the code where appropriate. * | |
6 | * * | |
7 | * Permission to use, copy, modify and distribute this software and its * | |
8 | * documentation strictly for non-commercialf purposes is hereby granted * | |
9 | * without fee, provided that the above copyright notice appears in all * | |
10 | * copies and that both the copyright notice and this permission notice * | |
11 | * appear in the supporting documentation. The authors make no claims * | |
12 | * about the suitability of this software for any purpose. It is * | |
13 | * provided "as is" without express or implied warranty. * | |
14 | **************************************************************************/ | |
1ee39b3a | 15 | |
16 | /* $Id: AliTRDpidRefMakerNN.cxx 27496 2008-07-22 08:35:45Z cblume $ */ | |
17 | ||
18 | //////////////////////////////////////////////////////////////////////////// | |
19 | // // | |
20 | // Builds the reference tree for the training of neural networks // | |
21 | // // | |
22 | //////////////////////////////////////////////////////////////////////////// | |
23 | ||
24 | #include "TSystem.h" | |
25 | #include "TDatime.h" | |
26 | #include "TPDGCode.h" | |
27 | #include "TH1F.h" | |
28 | #include "TH2F.h" | |
29 | #include "TFile.h" | |
30 | #include "TGraphErrors.h" | |
31 | #include "TTree.h" | |
32 | #include "TEventList.h" | |
33 | #include "TMultiLayerPerceptron.h" | |
34 | ||
35 | #include "AliPID.h" | |
36 | #include "AliESDtrack.h" | |
37 | #include "AliTrackReference.h" | |
38 | ||
39 | #include "AliTRDtrackV1.h" | |
1ee39b3a | 40 | #include "AliTRDpidUtil.h" |
41 | #include "AliTRDpidRefMakerNN.h" | |
42 | #include "AliTRDpidUtil.h" | |
43 | ||
da27d983 | 44 | #include "Cal/AliTRDCalPID.h" |
45 | #include "Cal/AliTRDCalPIDNN.h" | |
1ee39b3a | 46 | #include "info/AliTRDtrackInfo.h" |
47 | #include "info/AliTRDv0Info.h" | |
a5a3321d | 48 | #include "info/AliTRDpidInfo.h" |
1ee39b3a | 49 | |
50 | ClassImp(AliTRDpidRefMakerNN) | |
51 | ||
705f8b0a | 52 | //________________________________________________________________________ |
9ec5bce7 | 53 | AliTRDpidRefMakerNN::AliTRDpidRefMakerNN() |
705f8b0a | 54 | :AliTRDpidRefMaker() |
55 | ,fNet(NULL) | |
56 | ,fTrainMomBin(kAll) | |
57 | ,fEpochs(1000) | |
58 | ,fMinTrain(100) | |
59 | ,fDate(0) | |
60 | ,fDoTraining(0) | |
61 | ,fContinueTraining(0) | |
02281b83 | 62 | ,fTrainPath(0) |
705f8b0a | 63 | ,fScale(0) |
64 | ,fLy(0) | |
65 | ,fNtrkl(0) | |
66 | ,fRef(NULL) | |
67 | { | |
68 | // | |
69 | // Default constructor | |
70 | // | |
d80a6a00 | 71 | SetNameTitle("PIDrefMakerNN", "PID(NN) Reference Maker"); |
72 | ||
73 | memset(fTrain, 0, AliTRDCalPID::kNMom*sizeof(TEventList*)); | |
74 | memset(fTest, 0, AliTRDCalPID::kNMom*sizeof(TEventList*)); | |
75 | memset(fTrainData, 0, AliTRDCalPID::kNMom*sizeof(TTree*)); | |
76 | ||
77 | SetAbundance(.67); | |
78 | SetScaledEdx(Float_t(AliTRDCalPIDNN::kMLPscale)); | |
79 | TDatime datime; | |
80 | fDate = datime.GetDate(); | |
705f8b0a | 81 | } |
82 | ||
83 | //________________________________________________________________________ | |
84 | AliTRDpidRefMakerNN::AliTRDpidRefMakerNN(const char *name) | |
85 | :AliTRDpidRefMaker(name, "PID(NN) Reference Maker") | |
86 | ,fNet(NULL) | |
87 | ,fTrainMomBin(kAll) | |
88 | ,fEpochs(1000) | |
89 | ,fMinTrain(100) | |
90 | ,fDate(0) | |
91 | ,fDoTraining(0) | |
92 | ,fContinueTraining(0) | |
02281b83 | 93 | ,fTrainPath(0) |
705f8b0a | 94 | ,fScale(0) |
95 | ,fLy(0) | |
96 | ,fNtrkl(0) | |
97 | ,fRef(NULL) | |
1ee39b3a | 98 | { |
99 | // | |
100 | // Default constructor | |
101 | // | |
102 | ||
9ec5bce7 | 103 | memset(fTrain, 0, AliTRDCalPID::kNMom*sizeof(TEventList*)); |
104 | memset(fTest, 0, AliTRDCalPID::kNMom*sizeof(TEventList*)); | |
105 | memset(fTrainData, 0, AliTRDCalPID::kNMom*sizeof(TTree*)); | |
1ee39b3a | 106 | |
107 | SetAbundance(.67); | |
108 | SetScaledEdx(Float_t(AliTRDCalPIDNN::kMLPscale)); | |
109 | TDatime datime; | |
110 | fDate = datime.GetDate(); | |
1ee39b3a | 111 | } |
112 | ||
113 | ||
114 | //________________________________________________________________________ | |
115 | AliTRDpidRefMakerNN::~AliTRDpidRefMakerNN() | |
116 | { | |
117 | } | |
118 | ||
119 | ||
120 | //________________________________________________________________________ | |
705f8b0a | 121 | void AliTRDpidRefMakerNN::MakeTrainTestTrees() |
1ee39b3a | 122 | { |
9ec5bce7 | 123 | // Create output file and tree |
1ee39b3a | 124 | // Called once |
125 | ||
705f8b0a | 126 | fRef = new TFile("TRD.CalibPIDrefMakerNN.root", "RECREATE"); |
9ec5bce7 | 127 | for(Int_t ip = 0; ip < AliTRDCalPID::kNMom; ip++){ |
128 | fTrainData[ip] = new TTree(Form("fTrainData_%d", ip), Form("NN Reference Data for MomBin %d", ip)); | |
129 | fTrainData[ip] -> Branch("fdEdx", fdEdx, Form("fdEdx[%d]/F", AliTRDpidUtil::kNNslices)); | |
130 | fTrainData[ip] -> Branch("fPID", fPID, Form("fPID[%d]/F", AliPID::kSPECIES)); | |
131 | fTrainData[ip] -> Branch("fLy", &fLy, "fLy/I"); | |
132 | fTrainData[ip] -> Branch("fNtrkl", &fNtrkl, "fNtrkl/I"); | |
133 | ||
134 | fTrain[ip] = new TEventList(Form("fTrainMom%d", ip), Form("Training list for momentum intervall %d", ip)); | |
135 | fTest[ip] = new TEventList(Form("fTestMom%d", ip), Form("Test list for momentum intervall %d", ip)); | |
136 | } | |
1ee39b3a | 137 | } |
138 | ||
139 | ||
140 | ||
141 | //________________________________________________________________________ | |
142 | Bool_t AliTRDpidRefMakerNN::PostProcess() | |
143 | { | |
144 | // Draw result to the screen | |
145 | // Called once at the end of the query | |
146 | ||
d80a6a00 | 147 | TFile *fCalib = TFile::Open(Form("AnalysisResults.root")); |
9ec5bce7 | 148 | if (!fCalib) { |
149 | AliError("Calibration file not available"); | |
150 | return kFALSE; | |
151 | } | |
d80a6a00 | 152 | TDirectoryFile *dCalib = (TDirectoryFile*)fCalib->Get("TRD.CalibPIDrefMaker"); |
153 | if (!dCalib) { | |
154 | AliError("Calibration directory not available"); | |
155 | return kFALSE; | |
156 | } | |
157 | fData = (TTree*)dCalib->Get("RefPID"); | |
9ec5bce7 | 158 | if (!fData) { |
159 | AliError("Calibration data not available"); | |
160 | return kFALSE; | |
161 | } | |
705f8b0a | 162 | TObjArray *o = NULL; |
d80a6a00 | 163 | if(!(o = (TObjArray*)dCalib->Get("MonitorNN"))) { |
9ec5bce7 | 164 | AliWarning("Missing monitoring container."); |
165 | return kFALSE; | |
166 | } | |
167 | fContainer = (TObjArray*)o->Clone("monitor"); | |
168 | ||
d80a6a00 | 169 | |
9ec5bce7 | 170 | |
171 | if (!fRef) { | |
705f8b0a | 172 | AliDebug(2, "Loading file TRD.CalibPIDrefMakerNN.root"); |
173 | LoadFile("TRD.CalibPIDrefMakerNN.root"); | |
9ec5bce7 | 174 | } |
175 | else AliDebug(2, "file available"); | |
176 | ||
177 | if (!fRef) { | |
d80a6a00 | 178 | MakeTrainingSample(); |
9ec5bce7 | 179 | } |
180 | else AliDebug(2, "file available"); | |
181 | ||
182 | ||
183 | // build the training and the test list for the neural networks | |
184 | for(Int_t ip = 0; ip < AliTRDCalPID::kNMom; ip++){ | |
185 | MakeTrainingLists(ip); | |
186 | } | |
1ee39b3a | 187 | if(!fDoTraining) return kTRUE; |
188 | ||
9ec5bce7 | 189 | |
190 | ||
191 | // train the neural networks | |
1ee39b3a | 192 | gSystem->Exec(Form("mkdir ./Networks_%d/",fDate)); |
193 | AliDebug(2, Form("TrainMomBin [%d] [%d]", fTrainMomBin, kAll)); | |
194 | ||
195 | // train single network for a single momentum (recommended) | |
196 | if(!(fTrainMomBin == kAll)){ | |
9ec5bce7 | 197 | if(fTrain[fTrainMomBin] -> GetN() < fMinTrain){ |
198 | AliError("Warning in AliTRDpidRefMakerNN::PostProcess : Not enough events for training available! Please check Data sample!"); | |
1ee39b3a | 199 | return kFALSE; |
200 | } | |
201 | MakeRefs(fTrainMomBin); | |
1ee39b3a | 202 | MonitorTraining(fTrainMomBin); |
203 | } | |
204 | // train all momenta | |
205 | else{ | |
206 | for(Int_t iMomBin = 0; iMomBin < AliTRDCalPID::kNMom; iMomBin++){ | |
9ec5bce7 | 207 | if(fTrain[iMomBin] -> GetN() < fMinTrain){ |
7fe4e88b | 208 | AliError(Form("Warning in AliTRDpidRefMakerNN::PostProcess : Not enough events for training available for momentum bin [%d]! Please check Data sample!", iMomBin)); |
209 | continue; | |
1ee39b3a | 210 | } |
7fe4e88b | 211 | MakeRefs(iMomBin); |
1ee39b3a | 212 | MonitorTraining(iMomBin); |
213 | } | |
214 | } | |
215 | ||
216 | return kTRUE; // testing protection | |
217 | } | |
218 | ||
219 | ||
d80a6a00 | 220 | //________________________________________________________________________ |
221 | Bool_t AliTRDpidRefMakerNN::MakeTrainingSample() | |
222 | { | |
64d57299 | 223 | // convert AnalysisResults.root to training file |
d80a6a00 | 224 | TFile *fCalib = TFile::Open(Form("AnalysisResults.root")); |
225 | if (!fCalib) { | |
226 | AliError("Calibration file not available"); | |
227 | return kFALSE; | |
228 | } | |
229 | TDirectoryFile *dCalib = (TDirectoryFile*)fCalib->Get("TRD.CalibPIDrefMaker"); | |
230 | if (!dCalib) { | |
231 | AliError("Calibration directory not available"); | |
232 | return kFALSE; | |
233 | } | |
234 | fData = (TTree*)dCalib->Get("RefPID"); | |
235 | if (!fData) { | |
236 | AliError("Calibration data not available"); | |
237 | return kFALSE; | |
238 | } | |
239 | TObjArray *o = NULL; | |
240 | if(!(o = (TObjArray*)dCalib->Get("MonitorNN"))) { | |
241 | AliWarning("Missing monitoring container."); | |
242 | return kFALSE; | |
243 | } | |
244 | fContainer = (TObjArray*)o->Clone("monitor"); | |
245 | ||
246 | if (!fRef) { | |
247 | MakeTrainTestTrees(); | |
248 | ||
249 | // Convert the CaliPIDrefMaker tree to 11 (different momentum bin) trees for NN training | |
250 | ||
251 | LinkPIDdata(); | |
252 | for(Int_t ip=0; ip < AliTRDCalPID::kNMom; ip++){ | |
253 | for(Int_t is=0; is < AliPID::kSPECIES; is++) { | |
254 | memset(fPID, 0, AliPID::kSPECIES*sizeof(Float_t)); | |
255 | fPID[is] = 1; | |
256 | Int_t n(0); // index of data | |
257 | for(Int_t itrk = 0; itrk<fData->GetEntries() && n<kMaxStat; itrk++){ | |
258 | if(!(fData->GetEntry(itrk))) continue; | |
259 | if(fPIDdataArray->GetPID()!=is) continue; | |
260 | fNtrkl = fPIDdataArray->GetNtracklets(); | |
261 | for(Int_t ily=fPIDdataArray->GetNtracklets(); ily--;){ | |
262 | fLy = ily; | |
263 | if(fPIDdataArray->GetData(ily)->Momentum()!= ip) continue; | |
264 | memset(fdEdx, 0, AliTRDpidUtil::kNNslices*sizeof(Float_t)); | |
265 | for(Int_t islice=AliTRDCalPID::kNSlicesNN; islice--;){ | |
266 | fdEdx[islice]+=fPIDdataArray->GetData(ily)->fdEdx[islice]; | |
267 | fdEdx[islice]/=fScale; | |
268 | } | |
269 | fTrainData[ip] -> Fill(); | |
270 | n++; | |
271 | } | |
272 | } | |
273 | AliDebug(2, Form("%d %d %d", ip, is, n)); | |
274 | } | |
275 | } | |
276 | ||
277 | ||
278 | fRef -> cd(); | |
279 | for(Int_t ip = 0; ip < AliTRDCalPID::kNMom; ip++){ | |
280 | fTrainData[ip] -> Write(); | |
281 | } | |
282 | return kTRUE; | |
283 | ||
284 | } | |
285 | else AliWarning("Training file available. No conversion done!"); | |
286 | return kFALSE; | |
287 | ||
288 | } | |
289 | ||
1ee39b3a | 290 | //________________________________________________________________________ |
9ec5bce7 | 291 | void AliTRDpidRefMakerNN::MakeTrainingLists(Int_t mombin) |
1ee39b3a | 292 | { |
293 | // | |
294 | // build the training lists for the neural networks | |
295 | // | |
296 | ||
9ec5bce7 | 297 | if (!fRef) { |
298 | LoadFile(Form("TRD.Calib%s.root", GetName())); | |
1ee39b3a | 299 | } |
300 | ||
9ec5bce7 | 301 | if (!fRef) { |
302 | AliError("ERROR file for building training list not available"); | |
1ee39b3a | 303 | return; |
304 | } | |
305 | ||
9ec5bce7 | 306 | AliDebug(2, "\n Making training lists! \n"); |
1ee39b3a | 307 | |
9ec5bce7 | 308 | Int_t nPart[AliPID::kSPECIES]; |
309 | memset(nPart, 0, AliPID::kSPECIES*sizeof(Int_t)); | |
1ee39b3a | 310 | |
311 | // set needed branches | |
9ec5bce7 | 312 | fTrainData[mombin] -> SetBranchAddress("fdEdx", fdEdx); |
313 | fTrainData[mombin] -> SetBranchAddress("fPID", fPID); | |
314 | fTrainData[mombin] -> SetBranchAddress("fLy", &fLy); | |
315 | fTrainData[mombin] -> SetBranchAddress("fNtrkl", &fNtrkl); | |
316 | ||
1ee39b3a | 317 | // start first loop to check total number of each particle type |
9ec5bce7 | 318 | for(Int_t iEv=0; iEv < fTrainData[mombin] -> GetEntries(); iEv++){ |
319 | fTrainData[mombin] -> GetEntry(iEv); | |
1ee39b3a | 320 | |
321 | // use only events with goes through 6 layers TRD | |
9ec5bce7 | 322 | if(fNtrkl != AliTRDgeometry::kNlayer) continue; |
323 | ||
324 | if(fPID[AliPID::kElectron] == 1) | |
325 | nPart[AliPID::kElectron]++; | |
326 | else if(fPID[AliPID::kMuon] == 1) | |
327 | nPart[AliPID::kMuon]++; | |
328 | else if(fPID[AliPID::kPion] == 1) | |
329 | nPart[AliPID::kPion]++; | |
330 | else if(fPID[AliPID::kKaon] == 1) | |
331 | nPart[AliPID::kKaon]++; | |
332 | else if(fPID[AliPID::kProton] == 1) | |
333 | nPart[AliPID::kProton]++; | |
1ee39b3a | 334 | } |
335 | ||
336 | AliDebug(2, "Particle multiplicities:"); | |
9ec5bce7 | 337 | AliDebug(2, Form("Momentum[%d] Elecs[%d] Muons[%d] Pions[%d] Kaons[%d] Protons[%d]", mombin, nPart[AliPID::kElectron], nPart[AliPID::kMuon], nPart[AliPID::kPion], nPart[AliPID::kKaon], nPart[AliPID::kProton])); |
338 | ||
339 | ||
1ee39b3a | 340 | |
9ec5bce7 | 341 | // // implement counter of training and test sample size |
342 | Int_t iTrain = 0, iTest = 0; | |
1ee39b3a | 343 | |
344 | // set training sample size per momentum interval to 2/3 | |
345 | // of smallest particle counter and test sample to 1/3 | |
9ec5bce7 | 346 | iTrain = nPart[0]; |
347 | for(Int_t iPart = 1; iPart < AliPID::kSPECIES; iPart++){ | |
348 | // exclude muons and kaons if not availyable | |
349 | // this is neeeded since we do not have v0 candiates | |
350 | if((iPart == AliPID::kMuon || iPart == AliPID::kKaon) && (nPart[AliPID::kMuon] == 0 || nPart[AliPID::kKaon] == 0)) continue; | |
351 | if(iTrain > nPart[iPart]) | |
352 | iTrain = nPart[iPart]; | |
353 | } | |
354 | iTest = Int_t( iTrain * (1-fFreq)); | |
355 | iTrain = Int_t(iTrain * fFreq); | |
356 | AliDebug(2, Form("Momentum[%d] Train[%d] Test[%d]", mombin, iTrain, iTest)); | |
357 | ||
1ee39b3a | 358 | |
359 | ||
360 | // reset couters | |
9ec5bce7 | 361 | memset(nPart, 0, AliPID::kSPECIES*sizeof(Int_t)); |
1ee39b3a | 362 | |
363 | // start second loop to set the event lists | |
9ec5bce7 | 364 | for(Int_t iEv = 0; iEv < fTrainData[mombin] -> GetEntries(); iEv++){ |
365 | fTrainData[mombin] -> GetEntry(iEv); | |
366 | ||
367 | // set event list | |
368 | for(Int_t is = 0; is < AliPID::kSPECIES; is++){ | |
369 | if(nPart[is] < iTrain && fPID[is] == 1){ | |
370 | fTrain[mombin] -> Enter(iEv); | |
371 | nPart[is]++; | |
372 | } else if(nPart[is] < iTest+iTrain && fPID[is] == 1){ | |
373 | fTest[mombin] -> Enter(iEv); | |
374 | nPart[is]++; | |
1ee39b3a | 375 | } else continue; |
376 | } | |
377 | } | |
378 | ||
379 | AliDebug(2, "Particle multiplicities in both lists:"); | |
9ec5bce7 | 380 | AliDebug(2, Form("Momentum[%d] Elecs[%d] Muons[%d] Pions[%d] Kaons[%d] Protons[%d]", mombin, nPart[AliPID::kElectron], nPart[AliPID::kMuon], nPart[AliPID::kPion], nPart[AliPID::kKaon], nPart[AliPID::kProton])); |
381 | ||
382 | return; | |
1ee39b3a | 383 | } |
384 | ||
385 | ||
386 | //________________________________________________________________________ | |
387 | void AliTRDpidRefMakerNN::MakeRefs(Int_t mombin) | |
388 | { | |
389 | // | |
390 | // train the neural networks | |
391 | // | |
392 | ||
393 | ||
d80a6a00 | 394 | if (!fTrainData[mombin]) LoadFile(Form("TRD.CalibPIDrefMakerNN.root")); |
1ee39b3a | 395 | |
9ec5bce7 | 396 | if (!fTrainData[mombin]) { |
1ee39b3a | 397 | AliError("Tree for training list not available"); |
398 | return; | |
399 | } | |
400 | ||
401 | TDatime datime; | |
402 | fDate = datime.GetDate(); | |
403 | ||
404 | AliDebug(2, Form("Training momentum bin %d", mombin)); | |
405 | ||
406 | // set variable to monitor the training and to save the development of the networks | |
407 | Int_t nEpochs = fEpochs/kMoniTrain; | |
408 | AliDebug(2, Form("Training %d times %d epochs", kMoniTrain, nEpochs)); | |
409 | ||
410 | // make directories to save the networks | |
411 | gSystem->Exec(Form("rm -r ./Networks_%d/MomBin_%d",fDate, mombin)); | |
412 | gSystem->Exec(Form("mkdir ./Networks_%d/MomBin_%d",fDate, mombin)); | |
413 | ||
414 | // variable to check if network can load weights from previous training | |
9ec5bce7 | 415 | Bool_t bFirstLoop = kTRUE; |
1ee39b3a | 416 | |
417 | // train networks over several loops and save them after each loop | |
418 | for(Int_t iLoop = 0; iLoop < kMoniTrain; iLoop++){ | |
9ec5bce7 | 419 | fTrainData[mombin] -> SetEventList(fTrain[mombin]); |
420 | fTrainData[mombin] -> SetEventList(fTest[mombin]); | |
1ee39b3a | 421 | |
9ec5bce7 | 422 | AliDebug(2, Form("Momentum[%d] Trainingloop[%d]", mombin, iLoop)); |
1ee39b3a | 423 | |
9ec5bce7 | 424 | // check if network is already implemented |
425 | if(bFirstLoop == kTRUE){ | |
426 | fNet = new TMultiLayerPerceptron("fdEdx[0],fdEdx[1],fdEdx[2],fdEdx[3],fdEdx[4],fdEdx[5],fdEdx[6],fdEdx[7]:15:7:fPID[0],fPID[1],fPID[2],fPID[3],fPID[4]!",fTrainData[mombin],fTrain[mombin],fTest[mombin]); | |
427 | fNet -> SetLearningMethod(TMultiLayerPerceptron::kStochastic); // set learning method | |
428 | fNet -> TMultiLayerPerceptron::SetEta(0.001); // set learning speed | |
429 | if(!fContinueTraining){ | |
430 | if(AliLog::GetDebugLevel("","AliTRDpidRefMakerNN")>=2) fNet -> Train(nEpochs,"text update=10, graph"); | |
431 | else fNet -> Train(nEpochs,""); | |
1ee39b3a | 432 | } |
9ec5bce7 | 433 | else{ |
434 | fNet -> LoadWeights(Form("./Networks_%d/MomBin_%d/Net_%d",fTrainPath, mombin, kMoniTrain - 1)); | |
435 | if(AliLog::GetDebugLevel("","AliTRDpidRefMakerNN")>=2) fNet -> Train(nEpochs,"text update=10, graph+"); | |
436 | else fNet -> Train(nEpochs,"+"); | |
1ee39b3a | 437 | } |
9ec5bce7 | 438 | bFirstLoop = kFALSE; |
439 | } | |
440 | else{ | |
441 | if(AliLog::GetDebugLevel("","AliTRDpidRefMakerNN")>=2) fNet -> Train(nEpochs,"text update=10, graph+"); | |
442 | else fNet -> Train(nEpochs,"+"); | |
443 | } | |
1ee39b3a | 444 | |
9ec5bce7 | 445 | // save weights for monitoring of the training |
446 | fNet -> DumpWeights(Form("./Networks_%d/MomBin_%d/Net_%d",fDate, mombin, iLoop)); | |
1ee39b3a | 447 | } // end training loop |
448 | } | |
449 | ||
450 | ||
451 | ||
452 | //________________________________________________________________________ | |
453 | void AliTRDpidRefMakerNN::MonitorTraining(Int_t mombin) | |
454 | { | |
455 | // | |
456 | // train the neural networks | |
457 | // | |
458 | ||
d80a6a00 | 459 | if (!fTrainData[mombin]) LoadFile(Form("TRD.CalibPIDrefMakerNN.root")); |
9ec5bce7 | 460 | if (!fTrainData[mombin]) { |
1ee39b3a | 461 | AliError("Tree for training list not available"); |
462 | return; | |
463 | } | |
464 | ||
465 | // init networks and set event list | |
466 | for(Int_t iChamb = 0; iChamb < AliTRDgeometry::kNlayer; iChamb++){ | |
9ec5bce7 | 467 | fNet = new TMultiLayerPerceptron("fdEdx[0],fdEdx[1],fdEdx[2],fdEdx[3],fdEdx[4],fdEdx[5],fdEdx[6],fdEdx[7]:15:7:fPID[0],fPID[1],fPID[2],fPID[3],fPID[4]!",fTrainData[mombin],fTrain[mombin],fTest[mombin]); |
468 | fTrainData[mombin] -> SetEventList(fTrain[mombin]); | |
469 | fTrainData[mombin] -> SetEventList(fTest[mombin]); | |
1ee39b3a | 470 | } |
471 | ||
472 | // implement variables for likelihoods | |
473 | Float_t like[AliPID::kSPECIES][AliTRDgeometry::kNlayer]; | |
474 | memset(like, 0, AliPID::kSPECIES*AliTRDgeometry::kNlayer*sizeof(Float_t)); | |
475 | Float_t likeAll[AliPID::kSPECIES], totProb; | |
476 | ||
477 | Double_t pionEffiTrain[kMoniTrain], pionEffiErrTrain[kMoniTrain]; | |
478 | Double_t pionEffiTest[kMoniTrain], pionEffiErrTest[kMoniTrain]; | |
479 | memset(pionEffiTrain, 0, kMoniTrain*sizeof(Double_t)); | |
480 | memset(pionEffiErrTrain, 0, kMoniTrain*sizeof(Double_t)); | |
481 | memset(pionEffiTest, 0, kMoniTrain*sizeof(Double_t)); | |
482 | memset(pionEffiErrTest, 0, kMoniTrain*sizeof(Double_t)); | |
483 | ||
484 | // init histos | |
485 | const Float_t epsilon = 1/(2*(AliTRDpidUtil::kBins-1)); // get nice histos with bin center at 0 and 1 | |
486 | TH1F *hElecs, *hPions; | |
487 | hElecs = new TH1F("hElecs","Likelihood for electrons", AliTRDpidUtil::kBins, 0.-epsilon, 1.+epsilon); | |
488 | hPions = new TH1F("hPions","Likelihood for pions", AliTRDpidUtil::kBins, 0.-epsilon, 1.+epsilon); | |
489 | ||
9ec5bce7 | 490 | TGraphErrors *gEffisTrain = new TGraphErrors(kMoniTrain); |
491 | gEffisTrain -> SetLineColor(4); | |
492 | gEffisTrain -> SetMarkerColor(4); | |
493 | gEffisTrain -> SetMarkerStyle(29); | |
494 | gEffisTrain -> SetMarkerSize(1); | |
495 | ||
496 | TGraphErrors *gEffisTest = new TGraphErrors(kMoniTrain); | |
497 | gEffisTest -> SetLineColor(2); | |
498 | gEffisTest -> SetMarkerColor(2); | |
499 | gEffisTest -> SetMarkerStyle(29); | |
500 | gEffisTest -> SetMarkerSize(1); | |
1ee39b3a | 501 | |
502 | AliTRDpidUtil *util = new AliTRDpidUtil(); | |
503 | ||
504 | // monitor training progress | |
505 | for(Int_t iLoop = 0; iLoop < kMoniTrain; iLoop++){ | |
506 | ||
507 | // load weights | |
9ec5bce7 | 508 | fNet -> LoadWeights(Form("./Networks_%d/MomBin_%d/Net_%d",fDate, mombin, iLoop)); |
509 | AliDebug(2, Form("./Networks_%d/MomBin_%d/Net_%d",fDate, mombin, iLoop)); | |
1ee39b3a | 510 | |
511 | // event loop training list | |
1ee39b3a | 512 | |
9ec5bce7 | 513 | for(Int_t is = 0; is < AliPID::kSPECIES; is++){ |
1ee39b3a | 514 | |
9ec5bce7 | 515 | if(!((is == AliPID::kElectron) || (is == AliPID::kPion))) continue; |
1ee39b3a | 516 | |
9ec5bce7 | 517 | Int_t iChamb = 0; |
518 | // reset particle probabilities | |
1ee39b3a | 519 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ |
9ec5bce7 | 520 | likeAll[iPart] = 1./AliPID::kSPECIES; |
1ee39b3a | 521 | } |
9ec5bce7 | 522 | totProb = 0.; |
1ee39b3a | 523 | |
9ec5bce7 | 524 | AliDebug(2, Form("%d",fTrain[mombin] -> GetN())); |
525 | for(Int_t iEvent = 0; iEvent < fTrain[mombin] -> GetN(); iEvent++ ){ | |
526 | fTrainData[mombin] -> GetEntry(fTrain[mombin] -> GetEntry(iEvent)); | |
527 | // use event only if it is electron or pion | |
64d57299 | 528 | if(fPID[is] <1.e-5) continue; |
9ec5bce7 | 529 | // get the probabilities for each particle type in each chamber |
530 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
531 | like[iPart][iChamb] = fNet -> Result(fTrain[mombin] -> GetEntry(iEvent), iPart); | |
532 | likeAll[iPart] *= like[iPart][iChamb]; | |
533 | } | |
534 | //end chamber loop | |
535 | iChamb++; | |
536 | ||
537 | // get total probability and normalize it | |
538 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
539 | totProb += likeAll[iPart]; | |
540 | } | |
541 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
542 | likeAll[iPart] /= totProb; | |
543 | } | |
544 | ||
545 | if(iChamb == 5){ | |
546 | // fill likelihood distributions | |
547 | if(fPID[AliPID::kElectron] == 1) | |
548 | hElecs -> Fill(likeAll[AliPID::kElectron]); | |
549 | if(fPID[AliPID::kPion] == 1) | |
550 | hPions -> Fill(likeAll[AliPID::kElectron]); | |
551 | iChamb = 0; | |
552 | // reset particle probabilities | |
553 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
554 | likeAll[iPart] = 1./AliPID::kSPECIES; | |
555 | } | |
556 | totProb = 0.; | |
557 | } | |
558 | } // end event loop | |
559 | } // end species loop | |
560 | ||
1ee39b3a | 561 | |
562 | // calculate the pion efficiency and fill the graph | |
563 | util -> CalculatePionEffi(hElecs, hPions); | |
564 | pionEffiTrain[iLoop] = util -> GetPionEfficiency(); | |
565 | pionEffiErrTrain[iLoop] = util -> GetError(); | |
566 | ||
567 | gEffisTrain -> SetPoint(iLoop, iLoop+1, pionEffiTrain[iLoop]); | |
568 | gEffisTrain -> SetPointError(iLoop, 0, pionEffiErrTrain[iLoop]); | |
569 | hElecs -> Reset(); | |
570 | hPions -> Reset(); | |
571 | AliDebug(2, Form("TrainingLoop[%d] PionEfficiency[%f +/- %f]", iLoop, pionEffiTrain[iLoop], pionEffiErrTrain[iLoop])); | |
572 | // end training loop | |
573 | ||
574 | ||
575 | ||
9ec5bce7 | 576 | // monitor validation progress |
577 | for(Int_t is = 0; is < AliPID::kSPECIES; is++){ | |
1ee39b3a | 578 | |
9ec5bce7 | 579 | if(!((is == AliPID::kElectron) || (is == AliPID::kPion))) continue; |
580 | ||
581 | Int_t iChamb = 0; | |
1ee39b3a | 582 | // reset particle probabilities |
583 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
9ec5bce7 | 584 | likeAll[iPart] = 1./AliPID::kSPECIES; |
1ee39b3a | 585 | } |
586 | totProb = 0.; | |
587 | ||
9ec5bce7 | 588 | for(Int_t iEvent = 0; iEvent < fTest[mombin] -> GetN(); iEvent++ ){ |
589 | fTrainData[mombin] -> GetEntry(fTest[mombin] -> GetEntry(iEvent)); | |
590 | // use event only if it is electron or pion | |
64d57299 | 591 | if(TMath::Abs(fPID[is]- 1.)<1.e-5) continue; |
9ec5bce7 | 592 | |
593 | // get the probabilities for each particle type in each chamber | |
594 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
595 | like[iPart][iChamb] = fNet -> Result(fTest[mombin] -> GetEntry(iEvent), iPart); | |
596 | likeAll[iPart] *= like[iPart][iChamb]; | |
597 | } | |
598 | iChamb++; | |
599 | ||
600 | // get total probability and normalize it | |
601 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
602 | totProb += likeAll[iPart]; | |
603 | } | |
604 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
605 | likeAll[iPart] /= totProb; | |
606 | } | |
607 | ||
608 | if(iChamb == 5){ | |
609 | // fill likelihood distributions | |
610 | if(fPID[AliPID::kElectron] == 1) | |
611 | hElecs -> Fill(likeAll[AliPID::kElectron]); | |
612 | if(fPID[AliPID::kPion] == 1) | |
613 | hPions -> Fill(likeAll[AliPID::kElectron]); | |
614 | iChamb = 0; | |
615 | // reset particle probabilities | |
616 | for(Int_t iPart = 0; iPart < AliPID::kSPECIES; iPart++){ | |
617 | likeAll[iPart] = 1./AliPID::kSPECIES; | |
618 | } | |
619 | totProb = 0.; | |
620 | } | |
621 | } // end event loop | |
622 | } // end species loop | |
623 | ||
1ee39b3a | 624 | // calculate the pion efficiency and fill the graph |
625 | util -> CalculatePionEffi(hElecs, hPions); | |
626 | pionEffiTest[iLoop] = util -> GetPionEfficiency(); | |
627 | pionEffiErrTest[iLoop] = util -> GetError(); | |
628 | ||
629 | gEffisTest -> SetPoint(iLoop, iLoop+1, pionEffiTest[iLoop]); | |
630 | gEffisTest -> SetPointError(iLoop, 0, pionEffiErrTest[iLoop]); | |
631 | hElecs -> Reset(); | |
632 | hPions -> Reset(); | |
633 | AliDebug(2, Form("TestLoop[%d] PionEfficiency[%f +/- %f] \n", iLoop, pionEffiTest[iLoop], pionEffiErrTest[iLoop])); | |
634 | ||
9ec5bce7 | 635 | } // end validation loop |
1ee39b3a | 636 | |
637 | util -> Delete(); | |
638 | ||
639 | gEffisTest -> Draw("PAL"); | |
640 | gEffisTrain -> Draw("PL"); | |
641 | ||
642 | } | |
643 | ||
644 | ||
645 | //________________________________________________________________________ | |
9ec5bce7 | 646 | Bool_t AliTRDpidRefMakerNN::LoadFile(const Char_t *InFileNN) |
1ee39b3a | 647 | { |
648 | // | |
649 | // Loads the files and sets the event list | |
9ec5bce7 | 650 | // for neural network training. |
1ee39b3a | 651 | // Useable for training outside of the makeResults.C macro |
652 | // | |
653 | ||
9ec5bce7 | 654 | fRef = TFile::Open(Form("%s", InFileNN)); |
655 | if(!fRef) return 0; | |
656 | for(Int_t ip = 0; ip < AliTRDCalPID::kNMom; ip++){ | |
657 | fTrainData[ip] = (TTree*)fRef -> Get(Form("fTrainData_%d", ip)); | |
658 | } | |
1ee39b3a | 659 | |
660 | for(Int_t iMom = 0; iMom < AliTRDCalPID::kNMom; iMom++){ | |
9ec5bce7 | 661 | fTrain[iMom] = new TEventList(Form("fTrainMom%d", iMom), Form("Training list for momentum intervall %d", iMom)); |
662 | fTest[iMom] = new TEventList(Form("fTestMom%d", iMom), Form("Test list for momentum intervall %d", iMom)); | |
1ee39b3a | 663 | } |
9ec5bce7 | 664 | return 1; |
1ee39b3a | 665 | } |
666 | ||
667 |