1 | #include <cpgplot.h> |
---|
2 | #include <iostream> |
---|
3 | #include <math.h> |
---|
4 | #include <Utils/utils.hh> |
---|
5 | |
---|
6 | void findMinMax(const float *array, const int size, float &min, float &max) |
---|
7 | { |
---|
8 | min = max = array[0]; |
---|
9 | for(int i=1;i<size;i++) { |
---|
10 | if(array[i]<min) min=array[i]; |
---|
11 | if(array[i]>max) max=array[i]; |
---|
12 | } |
---|
13 | } |
---|
14 | |
---|
15 | float findMean(float *&array, int size) |
---|
16 | { |
---|
17 | float mean = array[0]; |
---|
18 | for(int i=1;i<size;i++) mean += array[i]; |
---|
19 | mean /= float(size); |
---|
20 | return mean; |
---|
21 | } |
---|
22 | |
---|
23 | float findStddev(float *&array, int size) |
---|
24 | { |
---|
25 | float mean = findMean(array,size); |
---|
26 | float stddev = (array[0]-mean) * (array[0]-mean); |
---|
27 | for(int i=1;i<size;i++) stddev += (array[i]-mean)*(array[i]-mean); |
---|
28 | stddev = sqrt(stddev/float(size-1)); |
---|
29 | return stddev; |
---|
30 | } |
---|
31 | |
---|
32 | float findMedian(float *&array, int size) |
---|
33 | { |
---|
34 | // NOTE: madfm = median absolute deviation from median |
---|
35 | float *newarray = new float[size]; |
---|
36 | float median; |
---|
37 | for(int i=0;i<size;i++) newarray[i] = array[i]; |
---|
38 | sort(newarray,0,size); |
---|
39 | if((size%2)==0) median = 0.5*(newarray[size/2-1]+newarray[size/2]); |
---|
40 | else median = newarray[size/2]; |
---|
41 | delete [] newarray; |
---|
42 | return median; |
---|
43 | } |
---|
44 | |
---|
45 | float findMADFM(float *&array, int size) |
---|
46 | { |
---|
47 | // NOTE: madfm = median absolute deviation from median |
---|
48 | float *newarray = new float[size]; |
---|
49 | float median = findMedian(array,size); |
---|
50 | float madfm; |
---|
51 | for(int i=0;i<size;i++) newarray[i] = fabs(array[i]-median); |
---|
52 | sort(newarray,0,size); |
---|
53 | if((size%2)==0) madfm = 0.5*(newarray[size/2-1]+newarray[size/2]); |
---|
54 | else madfm = newarray[size/2]; |
---|
55 | delete [] newarray; |
---|
56 | return madfm; |
---|
57 | } |
---|
58 | |
---|
59 | void findMedianStats(float *&array, int size, float &median, float &madfm) |
---|
60 | { |
---|
61 | // NOTE: madfm = median absolute deviation from median |
---|
62 | float *newarray = new float[size]; |
---|
63 | |
---|
64 | for(int i=0;i<size;i++) newarray[i] = array[i]; |
---|
65 | sort(newarray,0,size); |
---|
66 | if((size%2)==0) median = 0.5*(newarray[size/2-1]+newarray[size/2]); |
---|
67 | else median = newarray[size/2]; |
---|
68 | |
---|
69 | for(int i=0;i<size;i++) newarray[i] = fabs(array[i]-median); |
---|
70 | sort(newarray,0,size); |
---|
71 | if((size%2)==0) madfm = 0.5*(newarray[size/2-1]+newarray[size/2]); |
---|
72 | else madfm = newarray[size/2]; |
---|
73 | |
---|
74 | delete [] newarray; |
---|
75 | } |
---|
76 | |
---|
77 | void findMedianStats(float *&array, long size, float &median, float &madfm) |
---|
78 | { |
---|
79 | // NOTE: madfm = median absolute deviation from median |
---|
80 | float *newarray = new float[size]; |
---|
81 | |
---|
82 | for(int i=0;i<size;i++) newarray[i] = array[i]; |
---|
83 | sort(newarray,0,size); |
---|
84 | if((size%2)==0) median = 0.5*(newarray[size/2-1]+newarray[size/2]); |
---|
85 | else median = newarray[size/2]; |
---|
86 | |
---|
87 | for(int i=0;i<size;i++) newarray[i] = fabs(array[i]-median); |
---|
88 | sort(newarray,0,size); |
---|
89 | if((size%2)==0) madfm = 0.5*(newarray[size/2-1]+newarray[size/2]); |
---|
90 | else madfm = newarray[size/2]; |
---|
91 | |
---|
92 | delete [] newarray; |
---|
93 | } |
---|
94 | |
---|
95 | |
---|
96 | void findNormalStats(float *&array, int size, float &mean, float &stddev) |
---|
97 | { |
---|
98 | mean = array[0]; |
---|
99 | for(int i=1;i<size;i++){ |
---|
100 | mean += array[i]; |
---|
101 | } |
---|
102 | mean /= float(size); |
---|
103 | |
---|
104 | stddev = (array[0]-mean) * (array[0]-mean); |
---|
105 | for(int i=1;i<size;i++) stddev += (array[i]-mean)*(array[i]-mean); |
---|
106 | stddev = sqrt(stddev/float(size-1)); |
---|
107 | |
---|
108 | } |
---|
109 | |
---|
110 | void findTrimmedHistStatsOLD(float *array, const int size, float &tmean, float &tsigma) |
---|
111 | { |
---|
112 | |
---|
113 | const int nbin = 100; |
---|
114 | float *num = new float[nbin]; |
---|
115 | bool *keep = new bool[nbin]; |
---|
116 | |
---|
117 | // HOW MANY VALUES IN EACH BIN? |
---|
118 | float min,max; |
---|
119 | findMinMax(array,size,min,max); |
---|
120 | for(int i=0; i<nbin; i++) num[i]=0; |
---|
121 | for(int i=0; i<size; i++){ |
---|
122 | float fraction = (array[i] - min) / (max - min); |
---|
123 | int bin = (int)( floor(fraction*nbin) ); |
---|
124 | if((bin>=0)&&(bin<nbin)) num[bin]++; |
---|
125 | } |
---|
126 | int binmax = 0; |
---|
127 | for(int i=1; i<nbin; i++) if(num[i]>num[binmax]) binmax = i; |
---|
128 | for(int i=0; i<nbin; i++) keep[i] = (num[i]>=num[binmax]/2); |
---|
129 | float *newarray = new float[size]; |
---|
130 | int newsize = 0; |
---|
131 | for(int i=0; i<size; i++){ |
---|
132 | float fraction = (array[i] - min) / (max - min); |
---|
133 | int bin = (int)( floor(fraction*nbin) ); |
---|
134 | if(keep[bin]) newarray[newsize++] = array[i]; |
---|
135 | } |
---|
136 | // if(newsize==0) |
---|
137 | std::cerr << size << "<->" << newsize << std::endl; |
---|
138 | |
---|
139 | findNormalStats(newarray,newsize,tmean,tsigma); |
---|
140 | // cpgopen("tmp.ps/vps"); |
---|
141 | // cpghistlog(newsize,newarray,min,max,100,0); |
---|
142 | // cpghist(newsize,newarray,min,max,100,0); |
---|
143 | // cpgend(); |
---|
144 | delete [] num,keep,newarray; |
---|
145 | |
---|
146 | // tsigma *= trimToNormal; |
---|
147 | |
---|
148 | } |
---|
149 | void findTrimmedHistStats2(float *array, const int size, float &tmean, float &tsigma) |
---|
150 | { |
---|
151 | |
---|
152 | const int nbin = 200; |
---|
153 | float *num = new float[nbin]; |
---|
154 | bool *keep = new bool[nbin]; |
---|
155 | |
---|
156 | // HOW MANY VALUES IN EACH BIN? |
---|
157 | float min,max; |
---|
158 | findMinMax(array,size,min,max); |
---|
159 | for(int i=0; i<nbin; i++) num[i]=0; |
---|
160 | for(int i=0; i<size; i++){ |
---|
161 | float fraction = (array[i] - min) / (max - min); |
---|
162 | int bin = (int)( floor(fraction*nbin) ); |
---|
163 | if((bin>=0)&&(bin<nbin)) num[bin]++; |
---|
164 | } |
---|
165 | int binmax = 0; |
---|
166 | for(int i=1; i<nbin; i++) if(num[i]>num[binmax]) binmax = i; |
---|
167 | for(int i=0; i<nbin; i++) keep[i] = (num[i]>=num[binmax]/2); |
---|
168 | float *newarray = new float[size]; |
---|
169 | int newsize = 0; |
---|
170 | for(int i=0; i<size; i++){ |
---|
171 | float fraction = (array[i] - min) / (max - min); |
---|
172 | int bin = (int)( floor(fraction*nbin) ); |
---|
173 | if(keep[bin]) newarray[newsize++] = array[i]; |
---|
174 | } |
---|
175 | |
---|
176 | tmean = findMean(newarray,newsize); |
---|
177 | |
---|
178 | tsigma = newsize * (max-min)/float(nbin) / |
---|
179 | (num[binmax] * erf(sqrt(M_LN2)) * sqrt(2.*M_PI)); |
---|
180 | |
---|
181 | } |
---|
182 | |
---|
183 | void findTrimmedHistStats(float *array, const int size, float &tmean, float &tsigma) |
---|
184 | { |
---|
185 | float datamin,datamax; |
---|
186 | findMinMax(array,size,datamin,datamax); |
---|
187 | float *sorted = new float[size]; |
---|
188 | float *cumul = new float[size]; |
---|
189 | float *angle = new float[size]; |
---|
190 | for(int i=0;i<size;i++) sorted[i] = array[i]/(datamax-datamin); |
---|
191 | sort(sorted,0,size); |
---|
192 | for(int i=0;i<size;i++) cumul[i] = (float)i/(float)size; |
---|
193 | int width =(int)( 20. * log10((float)size)); |
---|
194 | for(int i=0;i<size;i++){ |
---|
195 | int beg,end; |
---|
196 | float slope,eSlope,inter,eInter,r; |
---|
197 | if(i<width/2) beg=0; else beg=i-width/2; |
---|
198 | if(i>=size-width/2) end=size-1; else end=i+width/2; |
---|
199 | if(linear_regression(size,sorted,cumul,beg,end,slope,eSlope,inter,eInter,r)==0) |
---|
200 | angle[i] = atan( fabs(slope) ) * 180. / M_PI; |
---|
201 | else angle[i] = 90.; |
---|
202 | } |
---|
203 | |
---|
204 | // int start = 0; |
---|
205 | // while(angle[start] < 45.) start++; |
---|
206 | // int finish = size-1; |
---|
207 | // while(angle[finish] < 45.) finish--; |
---|
208 | |
---|
209 | int maxpt = 0; |
---|
210 | for(int i = 1; i<size; i++) if(angle[i]>angle[maxpt]) maxpt=i; |
---|
211 | int start = maxpt; |
---|
212 | while((start>0)&&(angle[start]>45.)) start--; |
---|
213 | int finish = maxpt; |
---|
214 | while((finish < size-1)&&(angle[finish]>45.)) finish++; |
---|
215 | |
---|
216 | std::cerr << "npts = " << size << ", start = " << start << ", finish = " << finish << std::endl; |
---|
217 | |
---|
218 | int trimSize=0; |
---|
219 | float *newarray = new float[finish-start+1]; |
---|
220 | for(int i=0;i<finish-start+1;i++) newarray[i] = sorted[i+start]*(datamax-datamin); |
---|
221 | |
---|
222 | findNormalStats(newarray,finish-start+1,tmean,tsigma); |
---|
223 | |
---|
224 | } |
---|