1 | #include <cpgplot.h> |
---|
2 | #include <iostream> |
---|
3 | #include <math.h> |
---|
4 | #include <Utils/utils.hh> |
---|
5 | |
---|
6 | template <class T> T absval(T value) |
---|
7 | { |
---|
8 | if( value > 0) return value; |
---|
9 | else return 0-value; |
---|
10 | } |
---|
11 | //-------------------------------------------------------------------- |
---|
12 | |
---|
13 | template <class T> void findMinMax(const T *array, const int size, |
---|
14 | T &min, T &max) |
---|
15 | { |
---|
16 | min = max = array[0]; |
---|
17 | for(int i=1;i<size;i++) { |
---|
18 | if(array[i]<min) min=array[i]; |
---|
19 | if(array[i]>max) max=array[i]; |
---|
20 | } |
---|
21 | } |
---|
22 | template void findMinMax<int>(const int *array, const int size, |
---|
23 | int &min, int &max); |
---|
24 | template void findMinMax<float>(const float *array, const int size, |
---|
25 | float &min, float &max); |
---|
26 | //-------------------------------------------------------------------- |
---|
27 | |
---|
28 | template <class T> float findMean(T *array, int size) |
---|
29 | { |
---|
30 | float mean = array[0]; |
---|
31 | for(int i=1;i<size;i++) mean += array[i]; |
---|
32 | mean /= float(size); |
---|
33 | return mean; |
---|
34 | } |
---|
35 | template float findMean<int>(int *array, int size); |
---|
36 | template float findMean<float>(float *array, int size); |
---|
37 | //-------------------------------------------------------------------- |
---|
38 | |
---|
39 | template <class T> float findStddev(T *array, int size) |
---|
40 | { |
---|
41 | float mean = findMean(array,size); |
---|
42 | float stddev = (array[0]-mean) * (array[0]-mean); |
---|
43 | for(int i=1;i<size;i++) stddev += (array[i]-mean)*(array[i]-mean); |
---|
44 | stddev = sqrt(stddev/float(size-1)); |
---|
45 | return stddev; |
---|
46 | } |
---|
47 | template float findStddev<int>(int *array, int size); |
---|
48 | template float findStddev<float>(float *array, int size); |
---|
49 | //-------------------------------------------------------------------- |
---|
50 | |
---|
51 | template <class T> T findMedian(T *array, int size) |
---|
52 | { |
---|
53 | // NOTE: madfm = median absolute deviation from median |
---|
54 | T *newarray = new T[size]; |
---|
55 | T median; |
---|
56 | for(int i=0;i<size;i++) newarray[i] = array[i]; |
---|
57 | sort(newarray,0,size); |
---|
58 | if((size%2)==0) median = (newarray[size/2-1]+newarray[size/2])/2; |
---|
59 | else median = newarray[size/2]; |
---|
60 | delete [] newarray; |
---|
61 | return median; |
---|
62 | } |
---|
63 | template int findMedian<int>(int *array, int size); |
---|
64 | template float findMedian<float>(float *array, int size); |
---|
65 | //-------------------------------------------------------------------- |
---|
66 | |
---|
67 | template <class T> T findMADFM(T *array, int size) |
---|
68 | { |
---|
69 | // NOTE: madfm = median absolute deviation from median |
---|
70 | T *newarray = new T[size]; |
---|
71 | T median = findMedian(array,size); |
---|
72 | T madfm; |
---|
73 | for(int i=0;i<size;i++) newarray[i] = absval(array[i]-median); |
---|
74 | sort(newarray,0,size); |
---|
75 | if((size%2)==0) madfm = (newarray[size/2-1]+newarray[size/2])/2; |
---|
76 | else madfm = newarray[size/2]; |
---|
77 | delete [] newarray; |
---|
78 | return madfm; |
---|
79 | } |
---|
80 | template int findMADFM<int>(int *array, int size); |
---|
81 | template float findMADFM<float>(float *array, int size); |
---|
82 | //-------------------------------------------------------------------- |
---|
83 | |
---|
84 | template <class T> void findMedianStats(T *array, int size, |
---|
85 | T &median, T &madfm) |
---|
86 | { |
---|
87 | // NOTE: madfm = median absolute deviation from median |
---|
88 | if(size==0){ |
---|
89 | std::cerr << "Error in findMedianStats: zero sized array!\n"; |
---|
90 | return; |
---|
91 | } |
---|
92 | T *newarray = new T[size]; |
---|
93 | |
---|
94 | for(int i=0;i<size;i++) newarray[i] = array[i]; |
---|
95 | sort(newarray,0,size); |
---|
96 | if((size%2)==0) median = (newarray[size/2-1]+newarray[size/2])/2; |
---|
97 | else median = newarray[size/2]; |
---|
98 | |
---|
99 | for(int i=0;i<size;i++) newarray[i] = absval(array[i]-median); |
---|
100 | sort(newarray,0,size); |
---|
101 | if((size%2)==0) madfm = (newarray[size/2-1]+newarray[size/2])/2; |
---|
102 | else madfm = newarray[size/2]; |
---|
103 | |
---|
104 | delete [] newarray; |
---|
105 | } |
---|
106 | template void findMedianStats<int>(int *array, int size, |
---|
107 | int &median, int &madfm); |
---|
108 | template void findMedianStats<long>(long *array, int size, |
---|
109 | long &median, long &madfm); |
---|
110 | template void findMedianStats<float>(float *array, int size, |
---|
111 | float &median, float &madfm); |
---|
112 | template void findMedianStats<double>(double *array, int size, |
---|
113 | double &median, double &madfm); |
---|
114 | //-------------------------------------------------------------------- |
---|
115 | |
---|
116 | template <class T> void findMedianStats(T *array, int size, bool *isGood, |
---|
117 | T &median, T &madfm) |
---|
118 | { |
---|
119 | // NOTE: madfm = median absolute deviation from median |
---|
120 | int goodSize=0; |
---|
121 | for(int i=0;i<size;i++) if(isGood[i]) goodSize++; |
---|
122 | if(goodSize==0){ |
---|
123 | std::cerr << "Error in findMedianStats: no good values!\n"; |
---|
124 | return; |
---|
125 | } |
---|
126 | T *newarray = new T[goodSize]; |
---|
127 | for(int i=0;i<size;i++) if(isGood[i]) newarray[goodSize++] = array[i]; |
---|
128 | sort(newarray,0,goodSize); |
---|
129 | if((goodSize%2)==0) |
---|
130 | median = (newarray[goodSize/2-1]+newarray[goodSize/2])/2; |
---|
131 | else |
---|
132 | median = newarray[goodSize/2]; |
---|
133 | |
---|
134 | for(int i=0;i<goodSize;i++) newarray[i] = absval(newarray[i]-median); |
---|
135 | sort(newarray,0,goodSize); |
---|
136 | if((goodSize%2)==0) |
---|
137 | madfm = (newarray[goodSize/2-1]+newarray[goodSize/2])/2; |
---|
138 | else |
---|
139 | madfm = newarray[goodSize/2]; |
---|
140 | |
---|
141 | delete [] newarray; |
---|
142 | } |
---|
143 | template void findMedianStats<int>(int *array, int size, bool *isGood, |
---|
144 | int &median, int &madfm); |
---|
145 | template void findMedianStats<long>(long *array, int size, bool *isGood, |
---|
146 | long &median, long &madfm); |
---|
147 | template void findMedianStats<float>(float *array, int size, bool *isGood, |
---|
148 | float &median, float &madfm); |
---|
149 | template void findMedianStats<double>(double *array, int size, bool *isGood, |
---|
150 | double &median, double &madfm); |
---|
151 | //-------------------------------------------------------------------- |
---|
152 | |
---|
153 | |
---|
154 | template <class T> void findNormalStats(T *array, int size, |
---|
155 | float &mean, float &stddev) |
---|
156 | { |
---|
157 | if(size==0){ |
---|
158 | std::cerr << "Error in findNormalStats: zero sized array!\n"; |
---|
159 | return; |
---|
160 | } |
---|
161 | mean = array[0]; |
---|
162 | for(int i=1;i<size;i++){ |
---|
163 | mean += array[i]; |
---|
164 | } |
---|
165 | mean /= float(size); |
---|
166 | |
---|
167 | stddev = (array[0]-mean) * (array[0]-mean); |
---|
168 | for(int i=1;i<size;i++) stddev += (array[i]-mean)*(array[i]-mean); |
---|
169 | stddev = sqrt(stddev/float(size-1)); |
---|
170 | |
---|
171 | } |
---|
172 | template void findNormalStats<int>(int *array, int size, |
---|
173 | float &mean, float &stddev); |
---|
174 | template void findNormalStats<long>(long *array, int size, |
---|
175 | float &mean, float &stddev); |
---|
176 | template void findNormalStats<float>(float *array, int size, |
---|
177 | float &mean, float &stddev); |
---|
178 | template void findNormalStats<double>(double *array, int size, |
---|
179 | float &mean, float &stddev); |
---|
180 | //-------------------------------------------------------------------- |
---|
181 | |
---|
182 | template <class T> void findNormalStats(T *array, int size, bool *isGood, |
---|
183 | float &mean, float &stddev) |
---|
184 | { |
---|
185 | int goodSize=0; |
---|
186 | for(int i=0;i<size;i++) if(isGood[i]) goodSize++; |
---|
187 | if(goodSize==0){ |
---|
188 | std::cerr << "Error in findNormalStats: no good values!\n"; |
---|
189 | return; |
---|
190 | } |
---|
191 | int start=0; |
---|
192 | while(!isGood[start]){start++;} |
---|
193 | mean = array[start]; |
---|
194 | for(int i=start+1;i<size;i++){ |
---|
195 | if(isGood[i]) mean += array[i]; |
---|
196 | } |
---|
197 | mean /= float(goodSize); |
---|
198 | |
---|
199 | stddev = (array[start]-mean) * (array[start]-mean); |
---|
200 | for(int i=1;i<size;i++){ |
---|
201 | if(isGood[i]) stddev += (array[i]-mean)*(array[i]-mean); |
---|
202 | } |
---|
203 | stddev = sqrt(stddev/float(goodSize-1)); |
---|
204 | |
---|
205 | } |
---|
206 | template void findNormalStats<int>(int *array, int size, bool *isGood, |
---|
207 | float &mean, float &stddev); |
---|
208 | template void findNormalStats<long>(long *array, int size, bool *isGood, |
---|
209 | float &mean, float &stddev); |
---|
210 | template void findNormalStats<float>(float *array, int size, bool *isGood, |
---|
211 | float &mean, float &stddev); |
---|
212 | template void findNormalStats<double>(double *array, int size, bool *isGood, |
---|
213 | float &mean, float &stddev); |
---|
214 | //-------------------------------------------------------------------- |
---|
215 | |
---|
216 | void findTrimmedHistStatsOLD(float *array, const int size, |
---|
217 | float &tmean, float &tsigma) |
---|
218 | { |
---|
219 | |
---|
220 | const int nbin = 100; |
---|
221 | float *num = new float[nbin]; |
---|
222 | bool *keep = new bool[nbin]; |
---|
223 | |
---|
224 | // HOW MANY VALUES IN EACH BIN? |
---|
225 | float min,max; |
---|
226 | findMinMax(array,size,min,max); |
---|
227 | for(int i=0; i<nbin; i++) num[i]=0; |
---|
228 | for(int i=0; i<size; i++){ |
---|
229 | float fraction = (array[i] - min) / (max - min); |
---|
230 | int bin = (int)( floor(fraction*nbin) ); |
---|
231 | if((bin>=0)&&(bin<nbin)) num[bin]++; |
---|
232 | } |
---|
233 | int binmax = 0; |
---|
234 | for(int i=1; i<nbin; i++) if(num[i]>num[binmax]) binmax = i; |
---|
235 | for(int i=0; i<nbin; i++) keep[i] = (num[i]>=num[binmax]/2); |
---|
236 | float *newarray = new float[size]; |
---|
237 | int newsize = 0; |
---|
238 | for(int i=0; i<size; i++){ |
---|
239 | float fraction = (array[i] - min) / (max - min); |
---|
240 | int bin = (int)( floor(fraction*nbin) ); |
---|
241 | if(keep[bin]) newarray[newsize++] = array[i]; |
---|
242 | } |
---|
243 | |
---|
244 | findNormalStats(newarray,newsize,tmean,tsigma); |
---|
245 | delete [] num,keep,newarray; |
---|
246 | |
---|
247 | } |
---|
248 | //-------------------------------------------------------------------- |
---|
249 | |
---|
250 | void findTrimmedHistStats2(float *array, const int size, |
---|
251 | float &tmean, float &tsigma) |
---|
252 | { |
---|
253 | |
---|
254 | const int nbin = 200; |
---|
255 | float *num = new float[nbin]; |
---|
256 | bool *keep = new bool[nbin]; |
---|
257 | |
---|
258 | // HOW MANY VALUES IN EACH BIN? |
---|
259 | float min,max; |
---|
260 | findMinMax(array,size,min,max); |
---|
261 | for(int i=0; i<nbin; i++) num[i]=0; |
---|
262 | for(int i=0; i<size; i++){ |
---|
263 | float fraction = (array[i] - min) / (max - min); |
---|
264 | int bin = (int)( floor(fraction*nbin) ); |
---|
265 | if((bin>=0)&&(bin<nbin)) num[bin]++; |
---|
266 | } |
---|
267 | int binmax = 0; |
---|
268 | for(int i=1; i<nbin; i++) if(num[i]>num[binmax]) binmax = i; |
---|
269 | for(int i=0; i<nbin; i++) keep[i] = (num[i]>=num[binmax]/2); |
---|
270 | float *newarray = new float[size]; |
---|
271 | int newsize = 0; |
---|
272 | for(int i=0; i<size; i++){ |
---|
273 | float fraction = (array[i] - min) / (max - min); |
---|
274 | int bin = (int)( floor(fraction*nbin) ); |
---|
275 | if(keep[bin]) newarray[newsize++] = array[i]; |
---|
276 | } |
---|
277 | |
---|
278 | tmean = findMean(newarray,newsize); |
---|
279 | |
---|
280 | tsigma = newsize * (max-min)/float(nbin) / |
---|
281 | (num[binmax] * erf(sqrt(M_LN2)) * sqrt(2.*M_PI)); |
---|
282 | |
---|
283 | } |
---|
284 | //-------------------------------------------------------------------- |
---|
285 | |
---|
286 | void findTrimmedHistStats(float *array, const int size, |
---|
287 | float &tmean, float &tsigma) |
---|
288 | { |
---|
289 | float datamin,datamax; |
---|
290 | findMinMax(array,size,datamin,datamax); |
---|
291 | float *sorted = new float[size]; |
---|
292 | float *cumul = new float[size]; |
---|
293 | float *angle = new float[size]; |
---|
294 | for(int i=0;i<size;i++) sorted[i] = array[i]/(datamax-datamin); |
---|
295 | sort(sorted,0,size); |
---|
296 | for(int i=0;i<size;i++) cumul[i] = (float)i/(float)size; |
---|
297 | int width =(int)( 20. * log10((float)size)); |
---|
298 | for(int i=0;i<size;i++){ |
---|
299 | int beg,end; |
---|
300 | float slope,eSlope,inter,eInter,r; |
---|
301 | if(i<width/2) beg=0; else beg=i-width/2; |
---|
302 | if(i>=size-width/2) end=size-1; else end=i+width/2; |
---|
303 | if(linear_regression(size,sorted,cumul,beg,end,slope,eSlope,inter,eInter,r)==0) |
---|
304 | angle[i] = atan( fabs(slope) ) * 180. / M_PI; |
---|
305 | else angle[i] = 90.; |
---|
306 | } |
---|
307 | |
---|
308 | // int start = 0; |
---|
309 | // while(angle[start] < 45.) start++; |
---|
310 | // int finish = size-1; |
---|
311 | // while(angle[finish] < 45.) finish--; |
---|
312 | |
---|
313 | int maxpt = 0; |
---|
314 | for(int i = 1; i<size; i++) if(angle[i]>angle[maxpt]) maxpt=i; |
---|
315 | int start = maxpt; |
---|
316 | while((start>0)&&(angle[start]>45.)) start--; |
---|
317 | int finish = maxpt; |
---|
318 | while((finish < size-1)&&(angle[finish]>45.)) finish++; |
---|
319 | |
---|
320 | // std::cerr << "npts = " << size << ", start = " << start << ", finish = " << finish << std::endl; |
---|
321 | |
---|
322 | int trimSize=0; |
---|
323 | float *newarray = new float[finish-start+1]; |
---|
324 | for(int i=0;i<finish-start+1;i++) newarray[i] = sorted[i+start]*(datamax-datamin); |
---|
325 | |
---|
326 | findNormalStats(newarray,finish-start+1,tmean,tsigma); |
---|
327 | |
---|
328 | } |
---|