source: trunk/MagicSoft/Mars/mranforest/MRanTree.cc@ 7913

Last change on this file since 7913 was 7804, checked in by tbretz, 18 years ago
*** empty log message ***
File size: 22.7 KB
Line 
1/* ======================================================================== *\
2!
3! *
4! * This file is part of MARS, the MAGIC Analysis and Reconstruction
5! * Software. It is distributed to you in the hope that it can be a useful
6! * and timesaving tool in analysing Data of imaging Cerenkov telescopes.
7! * It is distributed WITHOUT ANY WARRANTY.
8! *
9! * Permission to use, copy, modify and distribute this software and its
10! * documentation for any purpose is hereby granted without fee,
11! * provided that the above copyright notice appear in all copies and
12! * that both that copyright notice and this permission notice appear
13! * in supporting documentation. It is provided "as is" without express
14! * or implied warranty.
15! *
16!
17!
18! Author(s): Thomas Hengstebeck 3/2003 <mailto:hengsteb@physik.hu-berlin.de>
19!
20! Copyright: MAGIC Software Development, 2000-2005
21!
22!
23\* ======================================================================== */
24
25/////////////////////////////////////////////////////////////////////////////
26//
27// MRanTree
28//
29// ParameterContainer for Tree structure
30//
31/////////////////////////////////////////////////////////////////////////////
32#include "MRanTree.h"
33
34#include <iostream>
35
36#include <TRandom.h>
37
38#include "MArrayI.h"
39#include "MArrayF.h"
40
41#include "MMath.h"
42
43#include "MLog.h"
44#include "MLogManip.h"
45
46ClassImp(MRanTree);
47
48using namespace std;
49
50
51// --------------------------------------------------------------------------
52// Default constructor.
53//
54MRanTree::MRanTree(const char *name, const char *title):fClassify(kTRUE),fNdSize(0), fNumTry(3)
55{
56
57 fName = name ? name : "MRanTree";
58 fTitle = title ? title : "Storage container for structure of a single tree";
59}
60
61// --------------------------------------------------------------------------
62// Copy constructor
63//
64MRanTree::MRanTree(const MRanTree &tree)
65{
66 fName = tree.fName;
67 fTitle = tree.fTitle;
68
69 fClassify = tree.fClassify;
70 fNdSize = tree.fNdSize;
71 fNumTry = tree.fNumTry;
72
73 fNumNodes = tree.fNumNodes;
74 fNumEndNodes = tree.fNumEndNodes;
75
76 fBestVar = tree.fBestVar;
77 fTreeMap1 = tree.fTreeMap1;
78 fTreeMap2 = tree.fTreeMap2;
79 fBestSplit = tree.fBestSplit;
80 fGiniDec = tree.fGiniDec;
81}
82
83void MRanTree::SetNdSize(Int_t n)
84{
85 // threshold nodesize of terminal nodes, i.e. the training data is splitted
86 // until there is only pure date in the subsets(=terminal nodes) or the
87 // subset size is LE n
88
89 fNdSize=TMath::Max(1,n);//at least 1 event per node
90}
91
92void MRanTree::SetNumTry(Int_t n)
93{
94 // number of trials in random split selection:
95 // choose at least 1 variable to split in
96
97 fNumTry=TMath::Max(1,n);
98}
99
100void MRanTree::GrowTree(TMatrix *mat, const MArrayF &hadtrue, const MArrayI &idclass,
101 MArrayI &datasort, const MArrayI &datarang, const MArrayF &tclasspop,
102 const Float_t &mean, const Float_t &square, const MArrayI &jinbag, const MArrayF &winbag,
103 const int nclass)
104{
105 // arrays have to be initialized with generous size, so number of total nodes (nrnodes)
106 // is estimated for worst case
107 const Int_t numdim =mat->GetNcols();
108 const Int_t numdata=winbag.GetSize();
109 const Int_t nrnodes=2*numdata+1;
110
111 // number of events in bootstrap sample
112 Int_t ninbag=0;
113 for (Int_t n=0;n<numdata;n++) if(jinbag[n]==1) ninbag++;
114
115 MArrayI bestsplit(nrnodes);
116 MArrayI bestsplitnext(nrnodes);
117
118 fBestVar.Set(nrnodes); fBestVar.Reset();
119 fTreeMap1.Set(nrnodes); fTreeMap1.Reset();
120 fTreeMap2.Set(nrnodes); fTreeMap2.Reset();
121 fBestSplit.Set(nrnodes); fBestSplit.Reset();
122 fGiniDec.Set(numdim); fGiniDec.Reset();
123
124
125 if(fClassify)
126 FindBestSplit=&MRanTree::FindBestSplitGini;
127 else
128 FindBestSplit=&MRanTree::FindBestSplitSigma;
129
130 // tree growing
131 BuildTree(datasort,datarang,hadtrue,idclass,bestsplit, bestsplitnext,
132 tclasspop,mean,square,winbag,ninbag,nclass);
133
134 // post processing, determine cut (or split) values fBestSplit
135 for(Int_t k=0; k<nrnodes; k++)
136 {
137 if (GetNodeStatus(k)==-1)
138 continue;
139
140 const Int_t &bsp =bestsplit[k];
141 const Int_t &bspn=bestsplitnext[k];
142 const Int_t &msp =fBestVar[k];
143
144 fBestSplit[k] = ((*mat)(bsp, msp)+(*mat)(bspn,msp))/2;
145 }
146
147 // resizing arrays to save memory
148 fBestVar.Set(fNumNodes);
149 fTreeMap1.Set(fNumNodes);
150 fTreeMap2.Set(fNumNodes);
151 fBestSplit.Set(fNumNodes);
152}
153
154int MRanTree::FindBestSplitGini(const MArrayI &datasort,const MArrayI &datarang,
155 const MArrayF &hadtrue,const MArrayI &idclass,
156 Int_t ndstart,Int_t ndend, const MArrayF &tclasspop,
157 const Float_t &mean, const Float_t &square, Int_t &msplit,
158 Float_t &decsplit,Int_t &nbest, const MArrayF &winbag,
159 const int nclass)
160{
161 const Int_t nrnodes = fBestSplit.GetSize();
162 const Int_t numdata = (nrnodes-1)/2;
163 const Int_t mdim = fGiniDec.GetSize();
164
165 // For the best split, msplit is the index of the variable (e.g Hillas par.,
166 // zenith angle ,...)
167 // split on. decsplit is the decreae in impurity measured by Gini-index.
168 // nsplit is the case number of value of msplit split on,
169 // and nsplitnext is the case number of the next larger value of msplit.
170
171 Int_t nbestvar=0;
172
173 // compute initial values of numerator and denominator of Gini-index,
174 // Gini index= pno/dno
175 Double_t pno=0;
176 Double_t pdo=0;
177
178 // tclasspop: sum of weights for events in class
179 for (Int_t j=0; j<nclass; j++) // loop over number of classes to classifiy
180 {
181 pno+=tclasspop[j]*tclasspop[j];
182 pdo+=tclasspop[j];
183 }
184
185 const Double_t crit0=pno/pdo; // weighted mean of weights
186
187 // start main loop through variables to find best split,
188 // (Gini-index as criterium crit)
189
190 Double_t critmax=-FLT_MAX;
191
192 // random split selection, number of trials = fNumTry
193 for (Int_t mt=0; mt<fNumTry; mt++) // we could try ALL variables???
194 {
195 const Int_t mvar= gRandom->Integer(mdim);
196 const Int_t mn = mvar*numdata;
197
198 // Gini index = rrn/rrd+rln/rld
199 Double_t rrn=pno;
200 Double_t rrd=pdo;
201 Double_t rln=0;
202 Double_t rld=0;
203
204 MArrayF wl(nclass); // left node //nclass
205 MArrayF wr(tclasspop); // right node//nclass
206
207 Double_t critvar=-FLT_MAX;
208 for(Int_t nsp=ndstart;nsp<=ndend-1;nsp++)
209 {
210 const Int_t &nc = datasort[mn+nsp];
211 const Int_t &k = idclass[nc];
212 const Float_t &u = winbag[nc];
213
214 // do classification, Gini index as split rule
215 rln +=u*(2*wl[k]+u); // += u*(wl[k]{i-1} + wl[k]{i-1}+u{i})
216 rld +=u; // sum of weights left from cut total
217 wl[k] +=u; // sum of weights left from cut for class k
218
219 rrn -=u*(2*wr[k]-u); // -= u*(wr[k]{i-1} + wr[k]{i-1}-u{i})
220 // rr0=0; rr0+=u*2*tclasspop[k]
221 // rrn = pno - rr0 + rln
222 rrd -=u; // sum of weights right from cut total
223 wr[k] -=u; // sum of weights right from cut for class k
224
225 // REPLACE BY?
226 // rr0 = 0
227 // rr0 += u*2*tclasspop[k]
228 // rrn = pno - rr0 + rln
229 // rrd = pdo - rld
230 // wr[k] = tclasspop[k] - wl[k]
231
232 // crit = (rln*(pdo - rld + 1) + pno - rr0) / rld*(pdo - rld)
233
234 /*
235 if (k==background)
236 continue;
237 crit = TMath::Max(MMath::SignificanceLiMa(rld, rld-wl[k]),
238 MMath::SignificanceLiMa(rrd, rrd-wr[k]))
239 */
240
241 // This condition is in fact a == (> cannot happen at all)
242 // This is because we cannot set the cut between two identical values
243 //if (datarang[mn+datasort[mn+nsp]]>=datarang[mn+datasort[mn+nsp+1]])
244 if (datarang[mn+nc]>=datarang[mn+datasort[mn+nsp+1]])
245 continue;
246
247 // If crit starts to become pretty large do WHAT???
248 //if (TMath::Min(rrd,rld)<=1.0e-5) // FIXME: CHECKIT FOR WEIGHTS!
249 // continue;
250
251 const Double_t crit=(rln/rld)+(rrn/rrd);
252 if (!TMath::Finite(crit))
253 continue;
254
255 // Search for the highest value of crit
256 if (crit<=critvar) continue;
257
258 // store the highest crit value and the corresponding event to cut at
259 nbestvar=nsp;
260 critvar=crit;
261 }
262
263 if (critvar<=critmax) continue;
264
265 msplit=mvar; // Variable in which to split
266 nbest=nbestvar; // event at which the best split was found
267 critmax=critvar;
268 }
269
270 // crit0 = MMath::SignificanceLiMa(pdo, pdo-tclasspop[0])
271 // mean increase of sensitivity
272 // decsplit = sqrt(critmax/crit0)
273 decsplit=critmax-crit0;
274
275 return critmax<-1.0e10 ? 1 : 0;
276}
277
278int MRanTree::FindBestSplitSigma(const MArrayI &datasort,const MArrayI &datarang,
279 const MArrayF &hadtrue, const MArrayI &idclass,
280 Int_t ndstart,Int_t ndend, const MArrayF &tclasspop,
281 const Float_t &mean, const Float_t &square, Int_t &msplit,
282 Float_t &decsplit,Int_t &nbest, const MArrayF &winbag,
283 const int nclass)
284{
285 const Int_t nrnodes = fBestSplit.GetSize();
286 const Int_t numdata = (nrnodes-1)/2;
287 const Int_t mdim = fGiniDec.GetSize();
288
289 // For the best split, msplit is the index of the variable (e.g Hillas par., zenith angle ,...)
290 // split on. decsplit is the decreae in impurity measured by Gini-index.
291 // nsplit is the case number of value of msplit split on,
292 // and nsplitnext is the case number of the next larger value of msplit.
293
294 Int_t nbestvar=0;
295
296 // compute initial values of numerator and denominator of split-index,
297
298 // resolution
299 //Double_t pno=-(tclasspop[0]*square-mean*mean)*tclasspop[0];
300 //Double_t pdo= (tclasspop[0]-1.)*mean*mean;
301
302 // n*resolution
303 //Double_t pno=-(tclasspop[0]*square-mean*mean)*tclasspop[0];
304 //Double_t pdo= mean*mean;
305
306 // variance
307 //Double_t pno=-(square-mean*mean/tclasspop[0]);
308 //Double_t pdo= (tclasspop[0]-1.);
309
310 // n*variance
311 Double_t pno= (square-mean*mean/tclasspop[0]);
312 Double_t pdo= 1.;
313
314 // 1./(n*variance)
315 //Double_t pno= 1.;
316 //Double_t pdo= (square-mean*mean/tclasspop[0]);
317
318 const Double_t crit0=pno/pdo;
319
320 // start main loop through variables to find best split,
321
322 Double_t critmin=FLT_MAX;
323
324 // random split selection, number of trials = fNumTry
325 for (Int_t mt=0; mt<fNumTry; mt++)
326 {
327 const Int_t mvar= gRandom->Integer(mdim);
328 const Int_t mn = mvar*numdata;
329
330 Double_t esumr =mean;
331 Double_t e2sumr=square;
332 Double_t esuml =0;
333 Double_t e2suml=0;
334
335 float wl=0.;// left node
336 float wr=tclasspop[0]; // right node
337
338 Double_t critvar=critmin;
339 for(Int_t nsp=ndstart;nsp<=ndend-1;nsp++)
340 {
341 const Int_t &nc=datasort[mn+nsp];
342 const Float_t &f=hadtrue[nc];;
343 const Float_t &u=winbag[nc];
344
345 e2suml+=u*f*f;
346 esuml +=u*f;
347 wl +=u;
348
349 //-------------------------------------------
350 // resolution
351 //const Double_t rln=(wl*e2suml-esuml*esuml)*wl;
352 //const Double_t rld=(wl-1.)*esuml*esuml;
353
354 // resolution times n
355 //const Double_t rln=(wl*e2suml-esuml*esuml)*wl;
356 //const Double_t rld=esuml*esuml;
357
358 // sigma
359 //const Double_t rln=(e2suml-esuml*esuml/wl);
360 //const Double_t rld=(wl-1.);
361
362 // sigma times n
363 Double_t rln=(e2suml-esuml*esuml/wl);
364 Double_t rld=1.;
365
366 // 1./(n*variance)
367 //const Double_t rln=1.;
368 //const Double_t rld=(e2suml-esuml*esuml/wl);
369 //-------------------------------------------
370
371 // REPLACE BY???
372 e2sumr-=u*f*f; // e2sumr = square - e2suml
373 esumr -=u*f; // esumr = mean - esuml
374 wr -=u; // wr = tclasspop[0] - wl
375
376 //-------------------------------------------
377 // resolution
378 //const Double_t rrn=(wr*e2sumr-esumr*esumr)*wr;
379 //const Double_t rrd=(wr-1.)*esumr*esumr;
380
381 // resolution times n
382 //const Double_t rrn=(wr*e2sumr-esumr*esumr)*wr;
383 //const Double_t rrd=esumr*esumr;
384
385 // sigma
386 //const Double_t rrn=(e2sumr-esumr*esumr/wr);
387 //const Double_t rrd=(wr-1.);
388
389 // sigma times n
390 const Double_t rrn=(e2sumr-esumr*esumr/wr);
391 const Double_t rrd=1.;
392
393 // 1./(n*variance)
394 //const Double_t rrn=1.;
395 //const Double_t rrd=(e2sumr-esumr*esumr/wr);
396 //-------------------------------------------
397
398 if (datarang[mn+nc]>=datarang[mn+datasort[mn+nsp+1]])
399 continue;
400
401 //if (TMath::Min(rrd,rld)<=1.0e-5)
402 // continue;
403
404 const Double_t crit=(rln/rld)+(rrn/rrd);
405 if (!TMath::Finite(crit))
406 continue;
407
408 if (crit>=critvar) continue;
409
410 nbestvar=nsp;
411 critvar=crit;
412 }
413
414 if (critvar>=critmin) continue;
415
416 msplit=mvar;
417 nbest=nbestvar;
418 critmin=critvar;
419 }
420
421 decsplit=crit0-critmin;
422
423 //return critmin>1.0e20 ? 1 : 0;
424 return decsplit<0 ? 1 : 0;
425}
426
427void MRanTree::MoveData(MArrayI &datasort,Int_t ndstart, Int_t ndend,
428 MArrayI &idmove,MArrayI &ncase,Int_t msplit,
429 Int_t nbest,Int_t &ndendl)
430{
431 // This is the heart of the BuildTree construction. Based on the best split
432 // the data in the part of datasort corresponding to the current node is moved to the
433 // left if it belongs to the left child and right if it belongs to the right child-node.
434 const Int_t numdata = ncase.GetSize();
435 const Int_t mdim = fGiniDec.GetSize();
436
437 MArrayI tdatasort(numdata);
438
439 // compute idmove = indicator of case nos. going left
440 for (Int_t nsp=ndstart;nsp<=ndend;nsp++)
441 {
442 const Int_t &nc=datasort[msplit*numdata+nsp];
443 idmove[nc]= nsp<=nbest?1:0;
444 }
445 ndendl=nbest;
446
447 // shift case. nos. right and left for numerical variables.
448 for(Int_t msh=0;msh<mdim;msh++)
449 {
450 Int_t k=ndstart-1;
451 for (Int_t n=ndstart;n<=ndend;n++)
452 {
453 const Int_t &ih=datasort[msh*numdata+n];
454 if (idmove[ih]==1)
455 tdatasort[++k]=datasort[msh*numdata+n];
456 }
457
458 for (Int_t n=ndstart;n<=ndend;n++)
459 {
460 const Int_t &ih=datasort[msh*numdata+n];
461 if (idmove[ih]==0)
462 tdatasort[++k]=datasort[msh*numdata+n];
463 }
464
465 for(Int_t m=ndstart;m<=ndend;m++)
466 datasort[msh*numdata+m]=tdatasort[m];
467 }
468
469 // compute case nos. for right and left nodes.
470
471 for(Int_t n=ndstart;n<=ndend;n++)
472 ncase[n]=datasort[msplit*numdata+n];
473}
474
475void MRanTree::BuildTree(MArrayI &datasort,const MArrayI &datarang, const MArrayF &hadtrue,
476 const MArrayI &idclass, MArrayI &bestsplit, MArrayI &bestsplitnext,
477 const MArrayF &tclasspop, const Float_t &tmean, const Float_t &tsquare, const MArrayF &winbag,
478 Int_t ninbag, const int nclass)
479{
480 // Buildtree consists of repeated calls to two void functions, FindBestSplit and MoveData.
481 // Findbestsplit does just that--it finds the best split of the current node.
482 // MoveData moves the data in the split node right and left so that the data
483 // corresponding to each child node is contiguous.
484 //
485 // buildtree bookkeeping:
486 // ncur is the total number of nodes to date. nodestatus(k)=1 if the kth node has been split.
487 // nodestatus(k)=2 if the node exists but has not yet been split, and =-1 if the node is
488 // terminal. A node is terminal if its size is below a threshold value, or if it is all
489 // one class, or if all the data-values are equal. If the current node k is split, then its
490 // children are numbered ncur+1 (left), and ncur+2(right), ncur increases to ncur+2 and
491 // the next node to be split is numbered k+1. When no more nodes can be split, buildtree
492 // returns.
493 const Int_t mdim = fGiniDec.GetSize();
494 const Int_t nrnodes = fBestSplit.GetSize();
495 const Int_t numdata = (nrnodes-1)/2;
496
497 MArrayI nodepop(nrnodes);
498 MArrayI nodestart(nrnodes);
499 MArrayI parent(nrnodes);
500
501 MArrayI ncase(numdata);
502 MArrayI idmove(numdata);
503 MArrayI iv(mdim);
504
505 MArrayF classpop(nrnodes*nclass);//nclass
506 MArrayI nodestatus(nrnodes);
507
508 for (Int_t j=0;j<nclass;j++)
509 classpop[j*nrnodes+0]=tclasspop[j];
510
511 MArrayF mean(nrnodes);
512 MArrayF square(nrnodes);
513 MArrayF lclasspop(tclasspop);
514
515 mean[0]=tmean;
516 square[0]=tsquare;
517
518
519 Int_t ncur=0;
520 nodepop[0]=ninbag;
521 nodestatus[0]=2;
522
523 // start main loop
524 for (Int_t kbuild=0; kbuild<nrnodes; kbuild++)
525 {
526 if (kbuild>ncur) break;
527 if (nodestatus[kbuild]!=2) continue;
528
529 // initialize for next call to FindBestSplit
530
531 const Int_t ndstart=nodestart[kbuild];
532 const Int_t ndend=ndstart+nodepop[kbuild]-1;
533
534 for (Int_t j=0;j<nclass;j++)
535 lclasspop[j]=classpop[j*nrnodes+kbuild];
536
537 Int_t msplit, nbest;
538 Float_t decsplit=0;
539
540 if ((this->*FindBestSplit)(datasort,datarang,hadtrue,idclass,ndstart,
541 ndend, lclasspop,mean[kbuild],square[kbuild],msplit,decsplit,
542 nbest,winbag,nclass))
543 {
544 nodestatus[kbuild]=-1;
545 continue;
546 }
547
548 fBestVar[kbuild]=msplit;
549 fGiniDec[msplit]+=decsplit;
550
551 bestsplit[kbuild]=datasort[msplit*numdata+nbest];
552 bestsplitnext[kbuild]=datasort[msplit*numdata+nbest+1];
553
554 Int_t ndendl;
555 MoveData(datasort,ndstart,ndend,idmove,ncase,
556 msplit,nbest,ndendl);
557
558 // leftnode no.= ncur+1, rightnode no. = ncur+2.
559 nodepop[ncur+1]=ndendl-ndstart+1;
560 nodepop[ncur+2]=ndend-ndendl;
561 nodestart[ncur+1]=ndstart;
562 nodestart[ncur+2]=ndendl+1;
563
564 // find class populations in both nodes
565 for (Int_t n=ndstart;n<=ndendl;n++)
566 {
567 const Int_t &nc=ncase[n];
568 const int j=idclass[nc];
569
570 // statistics left from cut
571 mean[ncur+1]+=hadtrue[nc]*winbag[nc];
572 square[ncur+1]+=hadtrue[nc]*hadtrue[nc]*winbag[nc];
573
574 // sum of weights left from cut
575 classpop[j*nrnodes+ncur+1]+=winbag[nc];
576 }
577
578 for (Int_t n=ndendl+1;n<=ndend;n++)
579 {
580 const Int_t &nc=ncase[n];
581 const int j=idclass[nc];
582
583 // statistics right from cut
584 mean[ncur+2] +=hadtrue[nc]*winbag[nc];
585 square[ncur+2]+=hadtrue[nc]*hadtrue[nc]*winbag[nc];
586
587 // sum of weights right from cut
588 classpop[j*nrnodes+ncur+2]+=winbag[nc];
589 }
590
591 // check on nodestatus
592
593 nodestatus[ncur+1]=2;
594 nodestatus[ncur+2]=2;
595 if (nodepop[ncur+1]<=fNdSize) nodestatus[ncur+1]=-1;
596 if (nodepop[ncur+2]<=fNdSize) nodestatus[ncur+2]=-1;
597
598
599 Double_t popt1=0;
600 Double_t popt2=0;
601 for (Int_t j=0;j<nclass;j++)
602 {
603 popt1+=classpop[j*nrnodes+ncur+1];
604 popt2+=classpop[j*nrnodes+ncur+2];
605 }
606
607 if(fClassify)
608 {
609 // check if only members of one class in node
610 for (Int_t j=0;j<nclass;j++)
611 {
612 if (classpop[j*nrnodes+ncur+1]==popt1) nodestatus[ncur+1]=-1;
613 if (classpop[j*nrnodes+ncur+2]==popt2) nodestatus[ncur+2]=-1;
614 }
615 }
616
617 fTreeMap1[kbuild]=ncur+1;
618 fTreeMap2[kbuild]=ncur+2;
619 parent[ncur+1]=kbuild;
620 parent[ncur+2]=kbuild;
621 nodestatus[kbuild]=1;
622 ncur+=2;
623 if (ncur>=nrnodes) break;
624 }
625
626 // determine number of nodes
627 fNumNodes=nrnodes;
628 for (Int_t k=nrnodes-1;k>=0;k--)
629 {
630 if (nodestatus[k]==0) fNumNodes-=1;
631 if (nodestatus[k]==2) nodestatus[k]=-1;
632 }
633
634 fNumEndNodes=0;
635 for (Int_t kn=0;kn<fNumNodes;kn++)
636 if(nodestatus[kn]==-1)
637 {
638 fNumEndNodes++;
639
640 Double_t pp=0;
641 for (Int_t j=0;j<nclass;j++)
642 {
643 if(classpop[j*nrnodes+kn]>pp)
644 {
645 // class + status of node kn coded into fBestVar[kn]
646 fBestVar[kn]=j-nclass;
647 pp=classpop[j*nrnodes+kn];
648 }
649 }
650
651 float sum=0;
652 for(int i=0;i<nclass;i++) sum+=classpop[i*nrnodes+kn];
653
654 fBestSplit[kn]=mean[kn]/sum;
655 }
656}
657
658Double_t MRanTree::TreeHad(const Float_t *evt)
659{
660 // to optimize on storage space node status and node class
661 // are coded into fBestVar:
662 // status of node kt = TMath::Sign(1,fBestVar[kt])
663 // class of node kt = fBestVar[kt]+2 (class defined by larger
664 // node population, actually not used)
665 // hadronness assigned to node kt = fBestSplit[kt]
666
667 // To get rid of the range check of the root classes
668 const Float_t *split = fBestSplit.GetArray();
669 const Int_t *map1 = fTreeMap1.GetArray();
670 const Int_t *map2 = fTreeMap2.GetArray();
671 const Int_t *best = fBestVar.GetArray();
672
673 Int_t kt=0;
674 for (Int_t k=0; k<fNumNodes; k++)
675 {
676 if (best[kt]<0)
677 break;
678
679 const Int_t m=best[kt];
680 kt = evt[m]<=split[kt] ? map1[kt] : map2[kt];
681 }
682
683 return split[kt];
684}
685
686Double_t MRanTree::TreeHad(const TVector &event)
687{
688 return TreeHad(event.GetMatrixArray());
689}
690
691Double_t MRanTree::TreeHad(const TMatrixFRow_const &event)
692{
693 return TreeHad(event.GetPtr());
694}
695
696Double_t MRanTree::TreeHad(const TMatrix &m, Int_t ievt)
697{
698#if ROOT_VERSION_CODE < ROOT_VERSION(4,00,8)
699 return TreeHad(TMatrixRow(m, ievt));
700#else
701 return TreeHad(TMatrixFRow_const(m, ievt));
702#endif
703}
704
705Bool_t MRanTree::AsciiWrite(ostream &out) const
706{
707 TString str;
708 Int_t k;
709
710 out.width(5);out<<fNumNodes<<endl;
711
712 for (k=0;k<fNumNodes;k++)
713 {
714 str=Form("%f",GetBestSplit(k));
715
716 out.width(5); out << k;
717 out.width(5); out << GetNodeStatus(k);
718 out.width(5); out << GetTreeMap1(k);
719 out.width(5); out << GetTreeMap2(k);
720 out.width(5); out << GetBestVar(k);
721 out.width(15); out << str<<endl;
722 out.width(5); out << GetNodeClass(k);
723 }
724 out<<endl;
725
726 return k==fNumNodes;
727}
Note: See TracBrowser for help on using the repository browser.