14.7 Do Two-Dimensional Distributions Differ? 645 111/j: decoding its row 1j=1-j*11; and column. mm=(m1=1i-k1)*(m2=1j-kj); pairs=tab[ki+1][kj+1]*tab[li+1][lj+1]; if (mm){ Not a tie. en1 +pairs; en2 +pairs; s +=(mm >0 pairs -pairs); Concordant,or discordant. else if (m1)en1 +=pairs; if (m2)en2 +pairs; 2 2 *tau=s/sqrt(en1*en2) 83g svar=(4.0*points+10.0)/(9.0*points*(points-1.0)); 19881992 *z=(*tau)/sqrt(svar); *prob=erfcc(fabs(*z)/1.4142136); 11800 (including this one) Cambridge NUMERICAL RECIPES CITED REFERENCES AND FURTHER READING: (Nort Lehmann,E.L.1975,Nonparametrics:Statistical Methods Based on Ranks (San Francisco: server computer University Press. THE Holden-Day). America to make one paper Downie,N.M.,and Heath,R.W.1965,Basic Statistica/Methods,2nd ed.(New York:Harper ART Row.pp.206-209. Norusis,M.J.1982,SPSS Introductory Guide:Basic Statistics and Operations;and 1985,SPSS- Programs X Advanced Statistics Guide (New York:McGraw-Hill). SCIENTIFIC 14.7 Do Two-Dimensional Distributions Differ? COMPUTING (ISBN 188810920 We here discuss a useful generalization of the K-S test (814.3)to two-dimensional distributions.This generalization is due to Fasano and Franceschini [1],a variant on an earlier Numerical Recipes 10-621 idea due to Peacock [2]. In a two-dimensional distribution,each data point is characterized by an (r,y)pair of idge.org 43108 values.An example near to our hearts is that each of the 19 neutrinos that were detected from Supernova 1987A is characterized by a time t and by an energy Ei(see [31).We might wish to know whether these measured pairs (E:),i =1...19 are consistent with a (outside theoretical model that predicts neutrino flux as a function of both time and energy-that is, a two-dimensional probability distribution in the (x,y)[here,(t,E)]plane.That would be North Software. a one-sample test.Or,given two sets of neutrino detections,from two comparable detectors, 6 we might want to know whether they are compatible with each other,a two-sample test. In the spirit of the tried-and-true,one-dimensional K-S test,we want to range over the (y)plane in search of some kind of maximum cumulative difference between two two-dimensional distributions.Unfortunately,cumulative probability distribution is not well-defined in more than one dimension!Peacock's insight was that a good surrogate is the integrated probability in each of four natural quadrants around a given point (,i), namely the total probabilities (or fraction of data)in (i,y>y),(yi), (i,y<yi),i,y<yi).The two-dimensional K-S statistic D is now taken to be the maximum difference(ranging both over data points and over quadrants)of the corresponding integrated probabilities.When comparing two data sets,the value of D may depend on which data set is ranged over.In that case,define an effective D as the average
14.7 Do Two-Dimensional Distributions Differ? 645 Permission is granted for internet users to make one paper copy for their own personal use. Further reproduction, or any copyin Copyright (C) 1988-1992 by Cambridge University Press. Programs Copyright (C) 1988-1992 by Numerical Recipes Software. Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) g of machinereadable files (including this one) to any server computer, is strictly prohibited. To order Numerical Recipes books or CDROMs, visit website http://www.nr.com or call 1-800-872-7423 (North America only), or send email to directcustserv@cambridge.org (outside North America). li=l/j; decoding its row lj=l-j*li; and column. mm=(m1=li-ki)*(m2=lj-kj); pairs=tab[ki+1][kj+1]*tab[li+1][lj+1]; if (mm) { Not a tie. en1 += pairs; en2 += pairs; s += (mm > 0 ? pairs : -pairs); Concordant, or discordant. } else { if (m1) en1 += pairs; if (m2) en2 += pairs; } } } *tau=s/sqrt(en1*en2); svar=(4.0*points+10.0)/(9.0*points*(points-1.0)); *z=(*tau)/sqrt(svar); *prob=erfcc(fabs(*z)/1.4142136); } CITED REFERENCES AND FURTHER READING: Lehmann, E.L. 1975, Nonparametrics: Statistical Methods Based on Ranks (San Francisco: Holden-Day). Downie, N.M., and Heath, R.W. 1965, Basic Statistical Methods, 2nd ed. (New York: Harper & Row), pp. 206–209. Norusis, M.J. 1982, SPSS Introductory Guide: Basic Statistics and Operations; and 1985, SPSSX Advanced Statistics Guide (New York: McGraw-Hill). 14.7 Do Two-Dimensional Distributions Differ? We here discuss a useful generalization of the K–S test (§14.3) to two-dimensional distributions. This generalization is due to Fasano and Franceschini [1], a variant on an earlier idea due to Peacock [2]. In a two-dimensional distribution, each data point is characterized by an (x, y) pair of values. An example near to our hearts is that each of the 19 neutrinos that were detected from Supernova 1987A is characterized by a time ti and by an energy Ei (see [3]). We might wish to know whether these measured pairs (ti, Ei), i = 1 ... 19 are consistent with a theoretical model that predicts neutrino flux as a function of both time and energy — that is, a two-dimensional probability distribution in the (x, y) [here, (t, E)] plane. That would be a one-sample test. Or, given two sets of neutrino detections, from two comparable detectors, we might want to know whether they are compatible with each other, a two-sample test. In the spirit of the tried-and-true, one-dimensional K–S test, we want to range over the (x, y) plane in search of some kind of maximum cumulative difference between two two-dimensional distributions. Unfortunately, cumulative probability distribution is not well-defined in more than one dimension! Peacock’s insight was that a good surrogate is the integrated probability in each of four natural quadrants around a given point (xi, yi), namely the total probabilities (or fraction of data) in (x>xi,y > yi), (x yi), (xxi,y < yi). The two-dimensional K–S statistic D is now taken to be the maximum difference (ranging both over data points and over quadrants) of the corresponding integrated probabilities. When comparing two data sets, the value of D may depend on which data set is ranged over. In that case, define an effective D as the average
646 Chapter 14.Statistical Description of Data 1256 6526 口 N △ of △ △ N △△0 △AA△4 83 ◇◇ △ DD 18881992 0 △ △△ 日△A D 11-600 (including this one) 872 …4△…次不2… 7423 from NUMERICAL RECIPES IN △ ◇ A △ △ 7 (North America to any server computer, tusers to make one paper /Cambridge University Press.Programs THE -2 是 ART △ .11.09 .12.09 strictly prohibited -3 -3 -2 -1 0 1 2 3 to dir Copyright (C) Figure 14.7.1.Two-dimensional distributions of 65 triangles and 35 squares.The two-dimensional K-S test finds that point one of whose quadrants (shown by dotted lines)maximizes the difference between fraction of triangles and fraction of squares.Then,equation (14.7.1)indicates whether the difference is ectcustser statistically significant,i.e.,whether the triangles and squares must have different underlying distributions. 1788-1982 OF SCIENTIFIC COMPUTING(ISBN v@cam of the two values obtained.If you are confused at this point about the exact definition of D. 10-621 don't fret;the accompanying computer routines amount to a precise algorithmic definition. Figure 14.7.1 gives a feeling for what is going on.The 65 triangles and 35 squares seem to have somewhat different distributions in the plane.The dotted lines are centered on the Numerical Recipes 43108-5 triangle that maximizes the D statistic,the maximum occurs in the upper-left quadrant.That quadrant contains only 0.12 of all the triangles,but it contains 0.56 of all the squares.The (outside value of D is thus 0.44.Is this statistically significant? Even for fixed sample sizes,it is unfortunately not rigorously true that the distribution of North Software. D in the null hypothesis is independent of the shape of the two-dimensional distribution.In this respect the two-dimensional K-S test is not as natural as its one-dimensional parent.However, extensive Monte Carlo integrations have shown that the distribution of the two-dimensional D is very nearly identical for even quite different distributions,as long as they have the same coefficient of correlation r,defined in the usual way by equation(14.5.1).In their paper, Fasano and Franceschini tabulate Monte Carlo results for (what amounts to)the distribution of D as a function of (of course)D,sample size N,and coefficient of correlation r.Analyzing their results,one finds that the significance levels for the two-dimensional K-S test can be summarized by the simple,though approximate,formulas, Probability (D>observed )=QKs VND (14.7.1) \1+√1-r2(0.25-0.75//N)
646 Chapter 14. Statistical Description of Data Permission is granted for internet users to make one paper copy for their own personal use. Further reproduction, or any copyin Copyright (C) 1988-1992 by Cambridge University Press. Programs Copyright (C) 1988-1992 by Numerical Recipes Software. Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) g of machinereadable files (including this one) to any server computer, is strictly prohibited. To order Numerical Recipes books or CDROMs, visit website http://www.nr.com or call 1-800-872-7423 (North America only), or send email to directcustserv@cambridge.org (outside North America). − 3 − 2 − 1 012 3 − 3 − 2 − 1 0 1 2 3 .11|.09 .65|.26 .12|.09 .12|.56 Figure 14.7.1. Two-dimensional distributions of 65 triangles and 35 squares. The two-dimensional K–S test finds that point one of whose quadrants (shown by dotted lines) maximizes the difference between fraction of triangles and fraction of squares. Then, equation (14.7.1) indicates whether the difference is statistically significant, i.e., whether the triangles and squares must have different underlying distributions. of the two values obtained. If you are confused at this point about the exact definition of D, don’t fret; the accompanying computer routines amount to a precise algorithmic definition. Figure 14.7.1 gives a feeling for what is going on. The 65 triangles and 35 squares seem to have somewhat different distributions in the plane. The dotted lines are centered on the triangle that maximizes the D statistic; the maximum occurs in the upper-left quadrant. That quadrant contains only 0.12 of all the triangles, but it contains 0.56 of all the squares. The value of D is thus 0.44. Is this statistically significant? Even for fixed sample sizes, it is unfortunately not rigorously true that the distribution of D in the null hypothesis is independent of the shape of the two-dimensional distribution. In this respect the two-dimensional K–S test is not as natural as its one-dimensional parent. However, extensive Monte Carlo integrations have shown that the distribution of the two-dimensional D is very nearly identical for even quite different distributions, as long as they have the same coefficient of correlation r, defined in the usual way by equation (14.5.1). In their paper, Fasano and Franceschini tabulate Monte Carlo results for (what amounts to) the distribution of D as a function of (of course) D, sample size N, and coefficient of correlation r. Analyzing their results, one finds that the significance levels for the two-dimensional K–S test can be summarized by the simple, though approximate, formulas, Probability (D > observed ) = QKS √N D 1 + √1 − r2(0.25 − 0.75/ √N) (14.7.1)
14.7 Do Two-Dimensional Distributions Differ? 647 for the one-sample case,and the same for the two-sample case,but with NIN2 N= (14.7.2) N1+N2 The above formulas are accurate enough when N 20,and when the indicated probability (significance level)is less than (more significant than)0.20 or so.When the indicated probability is 0.20,its value may not be accurate,but the implication that the data and model (or two data sets)are not significantly different is certainly correct.Notice that in the limit of r1 (perfect correlation),equations (14.7.1)and (14.7.2)reduce to equations (14.3.9)and (14.3.10):The two-dimensional data lie on a perfect straight line,and the two-dimensional K-S test becomes a one-dimensional K-S test. 5常 The significance level for the data in Figure 14.7.1,by the way,is about 0.001.This 8 establishes to a near-certainty that the triangles and squares were drawn from different distributions.(As in fact they were. Of course,if you do not want to rely on the Monte Carlo experiments embodied in equation (14.7.1),you can do your own:Generate a lot of synthetic data sets from your model,each one with the same number of points as the real data set.Compute D for each synthetic data set,using the accompanying computer routines (but ignoring their calculated probabilities).and count what fraction of the time these synthetic D's exceed the D from the real data.That fraction is your significance. One disadvantage of the two-dimensional tests,by comparison with their one-dimensional (North 令 progenitors,is that the two-dimensional tests require of order N2 operations:Two nested loops of order N take the place of an Nlog N sort.For small computers,this restricts the THE usefulness of the tests to N less than several thousand. We now give computer implementations.The one-sample case is embodied in the ART routine ks2d1s (that is,2-dimensions,1-sample).This routine calls a straightforward utility routine quadct to count points in the four quadrants,and it calls a user-supplied routine 9 Programs quadvl that must be capable of returning the integrated probability of an analytic model in each of four quadrants around an arbitrary (y)point.A trivial sample quadvl is shown; realistic quadvls can be quite complicated,often incorporating numerical quadratures over analytic two-dimensional distributions. to dir #include #include "nrutil.h" ectcustser OF SCIENTIFIC COMPUTING(ISBN void ks2dis(float x1,float yi0],unsigned long n1, 1881892 void (quadvl)(float,float,float *float *float *float *) float *d1,float *prob) Two-dimensional Kolmogorov-Smirnov test of one sample against a model.Given the x and y 10-621 coordinates of n1 data points in arrays x1[1..n1]and y1[1..n1],and given a user-supplied function quadvl that exemplifies the model,this routine returns the two-dimensional K-S statistic as d1,and its significance level as prob.Small values of prob show that the sample Numerical Recipes 43108 is significantly different from the model.Note that the test is slightly distribution-dependent, so prob is only an estimate. (outside void pearsn(float x[],float y[],unsigned long n,float r,float *prob, North Software. f0at*2】: float probks(float alam); void quadct(float x,float y,float xx[],float yy],unsigned long nn, Ame float *fa,float *fb,float *fc,float *fd); unsigned long ji float dum,dumm,fa,fb,fc,fd,ga,gb,gc,gd,r1,rr,sqen; *d1=0.0: for(j=1;j<n1;j++)[ Loop over the data points. quadct(x1[j],y1[j],x1,y1,n1,&fa,&fb,&fc,&fd); (*quadv1)(x1[j],y1[j],&ga,&gb,&gc,&gd); *d1=FMAX(*d1,fabs (fa-ga)); *d1=FMAX(*d1,fabs(fb-gb)); *d1=FMAX(*d1,fabs(fc-gc));
14.7 Do Two-Dimensional Distributions Differ? 647 Permission is granted for internet users to make one paper copy for their own personal use. Further reproduction, or any copyin Copyright (C) 1988-1992 by Cambridge University Press. Programs Copyright (C) 1988-1992 by Numerical Recipes Software. Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) g of machinereadable files (including this one) to any server computer, is strictly prohibited. To order Numerical Recipes books or CDROMs, visit website http://www.nr.com or call 1-800-872-7423 (North America only), or send email to directcustserv@cambridge.org (outside North America). for the one-sample case, and the same for the two-sample case, but with N = N1N2 N1 + N2 . (14.7.2) The above formulas are accurate enough when N >∼ 20, and when the indicated probability (significance level) is less than (more significant than) 0.20 or so. When the indicated probability is > 0.20, its value may not be accurate, but the implication that the data and model (or two data sets) are not significantly different is certainly correct. Notice that in the limit of r → 1 (perfect correlation), equations (14.7.1) and (14.7.2) reduce to equations (14.3.9) and (14.3.10): The two-dimensional data lie on a perfect straight line, and the two-dimensional K–S test becomes a one-dimensional K–S test. The significance level for the data in Figure 14.7.1, by the way, is about 0.001. This establishes to a near-certainty that the triangles and squares were drawn from different distributions. (As in fact they were.) Of course, if you do not want to rely on the Monte Carlo experiments embodied in equation (14.7.1), you can do your own: Generate a lot of synthetic data sets from your model, each one with the same number of points as the real data set. Compute D for each synthetic data set, using the accompanying computer routines (but ignoring their calculated probabilities), and count what fraction of the time these synthetic D’s exceed the D from the real data. That fraction is your significance. One disadvantage of the two-dimensional tests, by comparison with their one-dimensional progenitors, is that the two-dimensional tests require of order N2 operations: Two nested loops of order N take the place of an N log N sort. For small computers, this restricts the usefulness of the tests to N less than several thousand. We now give computer implementations. The one-sample case is embodied in the routine ks2d1s (that is, 2-dimensions, 1-sample). This routine calls a straightforward utility routine quadct to count points in the four quadrants, and it calls a user-supplied routine quadvl that must be capable of returning the integrated probability of an analytic model in each of four quadrants around an arbitrary (x, y) point. A trivial sample quadvl is shown; realistic quadvls can be quite complicated, often incorporating numerical quadratures over analytic two-dimensional distributions. #include #include "nrutil.h" void ks2d1s(float x1[], float y1[], unsigned long n1, void (*quadvl)(float, float, float *, float *, float *, float *), float *d1, float *prob) Two-dimensional Kolmogorov-Smirnov test of one sample against a model. Given the x and y coordinates of n1 data points in arrays x1[1..n1] and y1[1..n1], and given a user-supplied function quadvl that exemplifies the model, this routine returns the two-dimensional K-S statistic as d1, and its significance level as prob. Small values of prob show that the sample is significantly different from the model. Note that the test is slightly distribution-dependent, so prob is only an estimate. { void pearsn(float x[], float y[], unsigned long n, float *r, float *prob, float *z); float probks(float alam); void quadct(float x, float y, float xx[], float yy[], unsigned long nn, float *fa, float *fb, float *fc, float *fd); unsigned long j; float dum,dumm,fa,fb,fc,fd,ga,gb,gc,gd,r1,rr,sqen; *d1=0.0; for (j=1;j<=n1;j++) { Loop over the data points. quadct(x1[j],y1[j],x1,y1,n1,&fa,&fb,&fc,&fd); (*quadvl)(x1[j],y1[j],&ga,&gb,&gc,&gd); *d1=FMAX(*d1,fabs(fa-ga)); *d1=FMAX(*d1,fabs(fb-gb)); *d1=FMAX(*d1,fabs(fc-gc));
648 Chapter 14.Statistical Description of Data *d1=FMAX(*d1,fabs(fd-gd)); For both the sample and the model,the distribution is integrated in each of four quadrants,and the maximum difference is saved. pearsn(x1,y1,n1,&r1,&dum,&dumm); Get the linear correlation coefficient r1. sqen=sqrt((double)n1); rr=sqrt(1.0-r1*r1); Estimate the probability using the K-S probability function probks *prob=probks(*d1*sqen/(1.0+rr*(0.25-0.75/sqen))); http://www.nr. void quadct(float x,float y,float xx[,float yy],unsigned long nn, 83g float *fa,float *fb,float *fc,float *fd) Given an origin (x,y),and an array of nn points with coordinates xx[1..nn]and yy [1..nn] granted for 19881992 count how many of them are in each quadrant around the origin,and return the normalized fractions.Quadrants are labeled alphabetically,counterclockwise from the upper right.Used 100 by ks2d1s and ks2d2s. unsigned long k,na,nb,nc,nd; 7423 float ff; to to users to make one paper by Cambridge University Press. from NUMERICAL RECIPES IN C: na=nb=nc-nd=0; for (k=1;ky){ xx[k]>x？++na:++nb; (North America THE else xx[k]>x？++nd:++nc; ART Programs ff=1.0/nn; *fa-ff*na; send copy for thei *fb=ff*nb; *fc=ff*nc; *fd=ff*nd; email to directcustsen OF SCIENTIFIC COMPUTING(ISBN 0-521- #include "nrutil.h" void quadvl(float x,float y,float *fa,float *fb,float *fc,float *fd) v@cambr This is a sample of a user-supplied routine to be used with ks2d1s.In this case,the model distribution is uniform inside the square-1<x<1,-1 <y<1.In general this routine 1988-1992 by Numerical Recipes should return,for any point (x,y).the fraction of the total distribution in each of the four -43106 quadrants around that point.The fractions,fa,fb,fc,and fd,must add up to 1.Quadrants are alphabetical,counterclockwise from the upper right. (outside float qa,qb,qc,qd; Software. qa=FMIW(2.0,FNAX(0.0,1.0-x)); ying of qb=FHIW(2.0,FNAX(0.0,1.0-y)); qc=FMTW(2.0,FHAX(0.0,x+1.0)); qd=FMIN(2.0,FNAX(0.0,y+1.0)); *fa=0.25*ga*gb; *fb=0.25*qb*qc; *fc=0.25*qc*qd *fd=0.25*gd*ga; The routine ks2d2s is the two-sample case of the two-dimensional K-S test.It also calls quadct,pearsn,and probks.Being a two-sample test,it does not need an analytic model
648 Chapter 14. Statistical Description of Data Permission is granted for internet users to make one paper copy for their own personal use. Further reproduction, or any copyin Copyright (C) 1988-1992 by Cambridge University Press. Programs Copyright (C) 1988-1992 by Numerical Recipes Software. Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) g of machinereadable files (including this one) to any server computer, is strictly prohibited. To order Numerical Recipes books or CDROMs, visit website http://www.nr.com or call 1-800-872-7423 (North America only), or send email to directcustserv@cambridge.org (outside North America). *d1=FMAX(*d1,fabs(fd-gd)); For both the sample and the model, the distribution is integrated in each of four quadrants, and the maximum difference is saved. } pearsn(x1,y1,n1,&r1,&dum,&dumm); Get the linear correlation coefficient r1. sqen=sqrt((double)n1); rr=sqrt(1.0-r1*r1); Estimate the probability using the K-S probability function probks. *prob=probks(*d1*sqen/(1.0+rr*(0.25-0.75/sqen))); } void quadct(float x, float y, float xx[], float yy[], unsigned long nn, float *fa, float *fb, float *fc, float *fd) Given an origin (x, y), and an array of nn points with coordinates xx[1..nn] and yy[1..nn], count how many of them are in each quadrant around the origin, and return the normalized fractions. Quadrants are labeled alphabetically, counterclockwise from the upper right. Used by ks2d1s and ks2d2s. { unsigned long k,na,nb,nc,nd; float ff; na=nb=nc=nd=0; for (k=1;k y) { xx[k] > x ? ++na : ++nb; } else { xx[k] > x ? ++nd : ++nc; } } ff=1.0/nn; *fa=ff*na; *fb=ff*nb; *fc=ff*nc; *fd=ff*nd; } #include "nrutil.h" void quadvl(float x, float y, float *fa, float *fb, float *fc, float *fd) This is a sample of a user-supplied routine to be used with ks2d1s. In this case, the model distribution is uniform inside the square −1 <x< 1, −1 <y< 1. In general this routine should return, for any point (x, y), the fraction of the total distribution in each of the four quadrants around that point. The fractions, fa, fb, fc, and fd, must add up to 1. Quadrants are alphabetical, counterclockwise from the upper right. { float qa,qb,qc,qd; qa=FMIN(2.0,FMAX(0.0,1.0-x)); qb=FMIN(2.0,FMAX(0.0,1.0-y)); qc=FMIN(2.0,FMAX(0.0,x+1.0)); qd=FMIN(2.0,FMAX(0.0,y+1.0)); *fa=0.25*qa*qb; *fb=0.25*qb*qc; *fc=0.25*qc*qd; *fd=0.25*qd*qa; } The routine ks2d2s is the two-sample case of the two-dimensional K–S test. It also calls quadct, pearsn, and probks. Being a two-sample test, it does not need an analytic model
14.7 Do Two-Dimensional Distributions Differ? 649 #include #include "nrutil.h" void ks2d2s(f1oatx10,f1oaty1☐，unsigned1ongn1,f1oatx2☐，f1oaty2[0, unsigned long n2,float *d,float *prob) Two-dimensional Kolmogorov-Smirnov test on two samples.Given the x and y coordinates of the first sample as n1 values in arrays x1[1..n1]and y1[1..n1],and likewise for the second sample,n2 values in arrays x2 and y2,this routine returns the two-dimensional,two-sample K-S statistic as d,and its significance level as prob.Small values of prob show that the two samples are significantly different.Note that the test is slightly distribution-dependent,so prob is only an estimate. void pearsn(float x[],float y[],unsigned long n,float *r,float +prob, 5常 float *z); float probks(float alam); void quadct(float x,float y,float xx[],float yy[],unsigned long nn, 18881892 float *fa,float *fb,float *fc,float *fd); unsigned long j; float di,d2,dum,dumm,fa,fb,fc,fd,ga,gb,gc,gd,r1,r2,rr,sqen; d1=0.0; from NUMERICAL RECIPES I for(j=1:j<=n1:j++){ First,use points in the first sample as ori- quadct(x1[j],y1[j],x1,y1,n1,&fa,&fb,&fc,&fd);gins. quadct (x1[j],y1[j],x2,y2,n2,&ga,&gb,&gc,&gd); d1=FMAX(d1,fabs(fa-ga)); d1=FMAX(d1,fabs(fb-gb)); d1=FMAX(d1,fabs(fc-gc)); computer, Americ Press. d1=FMAX(d1,fabs(fd-gd)); ART d2=0.0; for(j=1;j<=n2;j++)[ Then,use points in the second sample as Programs quadct(x2[j],y2[j],x1,y1,n1,&fa,&fb,&fc,&fd); origins. quadct(x2[j],y2[j],x2,y2,n2,&ga,&gb,&gc,&gd); d2=FMAX(d2,fabs(fa-ga)); SCIENTIFIC d2=FMAX(d2,fabs(fb-gb)); d2=FMAX(d2,fabs(fc-gc)); to dir d2=FMAX(d2,fabs(fd-gd)); *d=0.5*(d1+d2); Average the K-S statistics sqen=sqrt(n1*n2/(double)(n1+n2)); 198918920 COMPUTING(ISBN pearsn(x1,y1,n1,&r1,&dum,&dumm); Get the linear correlation coefficient for each pearsn(x2,y2,n2,&r2,&dum,&dumm); sample. rr=sqrt(1.0-0.5*(r1*r1+r2*r2)); Estimate the probability using the K-S probability function probks. *prob=probks(*d*sqen/(1.0+rr*(0.25-0.75/sqen))); Numerical Recipes -43108 (outside 膜 Software. CITED REFERENCES AND FURTHER READING: Fasano.G.and Franceschini,A.1987,Monthly Notices of the Royal Astronomical Society. Ame vol.225,Pp.155-170.[1] Peacock,J.A.1983,Monthly Notices of the Royal Astronomical Society,vol.202,pp.615-627.[2] Spergel,D.N.,Piran,T.,Loeb,A.,Goodman,J.,and Bahcall,J.N.1987,Science,vol.237. pp.1471-1473.3
14.7 Do Two-Dimensional Distributions Differ? 649 Permission is granted for internet users to make one paper copy for their own personal use. Further reproduction, or any copyin Copyright (C) 1988-1992 by Cambridge University Press. Programs Copyright (C) 1988-1992 by Numerical Recipes Software. Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) g of machinereadable files (including this one) to any server computer, is strictly prohibited. To order Numerical Recipes books or CDROMs, visit website http://www.nr.com or call 1-800-872-7423 (North America only), or send email to directcustserv@cambridge.org (outside North America). #include #include "nrutil.h" void ks2d2s(float x1[], float y1[], unsigned long n1, float x2[], float y2[], unsigned long n2, float *d, float *prob) Two-dimensional Kolmogorov-Smirnov test on two samples. Given the x and y coordinates of the first sample as n1 values in arrays x1[1..n1] and y1[1..n1], and likewise for the second sample, n2 values in arrays x2 and y2, this routine returns the two-dimensional, two-sample K-S statistic as d, and its significance level as prob. Small values of prob show that the two samples are significantly different. Note that the test is slightly distribution-dependent, so prob is only an estimate. { void pearsn(float x[], float y[], unsigned long n, float *r, float *prob, float *z); float probks(float alam); void quadct(float x, float y, float xx[], float yy[], unsigned long nn, float *fa, float *fb, float *fc, float *fd); unsigned long j; float d1,d2,dum,dumm,fa,fb,fc,fd,ga,gb,gc,gd,r1,r2,rr,sqen; d1=0.0; for (j=1;j<=n1;j++) { First, use points in the first sample as oriquadct(x1[j],y1[j],x1,y1,n1,&fa,&fb,&fc,&fd); gins. quadct(x1[j],y1[j],x2,y2,n2,&ga,&gb,&gc,&gd); d1=FMAX(d1,fabs(fa-ga)); d1=FMAX(d1,fabs(fb-gb)); d1=FMAX(d1,fabs(fc-gc)); d1=FMAX(d1,fabs(fd-gd)); } d2=0.0; for (j=1;j<=n2;j++) { Then, use points in the second sample as quadct(x2[j],y2[j],x1,y1,n1,&fa,&fb,&fc,&fd); origins. quadct(x2[j],y2[j],x2,y2,n2,&ga,&gb,&gc,&gd); d2=FMAX(d2,fabs(fa-ga)); d2=FMAX(d2,fabs(fb-gb)); d2=FMAX(d2,fabs(fc-gc)); d2=FMAX(d2,fabs(fd-gd)); } *d=0.5*(d1+d2); Average the K-S statistics. sqen=sqrt(n1*n2/(double)(n1+n2)); pearsn(x1,y1,n1,&r1,&dum,&dumm); Get the linear correlation coefficient for each pearsn(x2,y2,n2,&r2,&dum,&dumm); sample. rr=sqrt(1.0-0.5*(r1*r1+r2*r2)); Estimate the probability using the K-S probability function probks. *prob=probks(*d*sqen/(1.0+rr*(0.25-0.75/sqen))); } CITED REFERENCES AND FURTHER READING: Fasano, G. and Franceschini, A. 1987, Monthly Notices of the Royal Astronomical Society, vol. 225, pp. 155–170. [1] Peacock, J.A. 1983, Monthly Notices of the Royal Astronomical Society, vol. 202, pp. 615–627. [2] Spergel, D.N., Piran, T., Loeb, A., Goodman, J., and Bahcall, J.N. 1987, Science, vol. 237, pp. 1471–1473. [3]
650 Chapter 14.Statistical Description of Data 14.8 Savitzky-Golay Smoothing Filters In $13.5 we learned something about the construction and application of digital filters, but little guidance was given on which particular filter to use.That,of course,depends on what you want to accomplish by filtering.One obvious use for low-pass filters is to smooth noisy data. The premise of data smoothing is that one is measuring a variable that is both slowly varying and also corrupted by random noise.Then it can sometimes be useful to replace each data point by some kind of local average of surrounding data points.Since nearby points measure very nearly the same underlying value,averaging can reduce the level of noise without (much)biasing the value obtained. 8 We must comment editorially that the smoothing of data lies in a murky area,beyond the fringe of some better posed,and therefore more highly recommended,techniques that are discussed elsewhere in this book.If you are fitting data to a parametric model,for example (see Chapter 15),it is almost always better to use raw data than to use data that has been pre-processed by a smoothing procedure.Another alternative to blind smoothing is so-called optimal"or Wiener filtering,as discussed in $13.3 and more generally in $13.6.Data smoothing is probably most justified when it is used simply as a graphical technique,to guide the eye through a forest of data points all with large error bars;or as a means of making initial rough estimates of simple parameters from a graph. 令 In this section we discuss a particular type of low-pass filter,well-adapted for data smoothing,and termed variously Savitzky-Golay[1],least-squares [2],or DISPO (Digital Smoothing Polynomial)[3]filters.Rather than having their properties defined in the Fourier domain,and then translated to the time domain,Savitzky-Golay filters derive directly from 、 Press. a particular formulation of the data smoothing problem in the time domain,as we will now see.Savitzky-Golay filters were initially (and are still often)used to render visible the relative widths and heights of spectral lines in noisy spectrometric data. Recall that a digital filter is applied to a series of equally spaced data values fa=f(t:), whereti=to +i for some constant sample spacing A and i =...-2,-1,0,1,2,.... SCIENTIFIC We have seen (813.5)that the simplest type of digital filter (the nonrecursive or finite impulse response filter)replaces each data value fi by a linear combination gi of itself and some number of nearby neighbors. nR 9= Cnfitn (14.8.1) n=-nL Here n is the number of points used"to the left"of a data point i,i.e.,earlier than it,while nR is the number used to the right,i.e.,later.A so-called causal filter would have n =0. 10621 As a starting point for understanding Savitzky-Golay filters,consider the simplest possible averaging procedure:For some fixed nL=nR,compute each g:as the average of 43106 the data points from fto fThis is sometimes called moving window averaging Numerical Recipes and corresponds to equation (14.8.1)with constant c 1/(nL +nR+1).If the underlying function is constant,or is changing linearly with time (increasing or decreasing),then no (outside bias is introduced into the result.Higher points at one end of the averaging interval are on the average balanced by lower points at the other end.A bias is introduced,however,if 首 Software. the underlying function has a nonzero second derivative.At a local maximum,for example, moving window averaging always reduces the function value.In the spectrometric application, a narrow spectral line has its height reduced and its width increased.Since these parameters are themselves of physical interest,the bias introduced is distinctly undesirable. Note,however,that moving window averaging does preserve the area under a spectral line,which is its zeroth moment,and also (if the window is symmetric with n=nR)its mean position in time,which is its first moment.What is violated is the second moment. equivalent to the line width. The idea of Savitzky-Golay filtering is to find filter coefficients cn that preserve higher moments.Equivalently,the idea is to approximate the underlying function within the moving window not by a constant (whose estimate is the average),but by a polynomial of higher order,typically quadratic or quartic:For each point fi,we least-squares fit a polynomial to all
650 Chapter 14. Statistical Description of Data Permission is granted for internet users to make one paper copy for their own personal use. Further reproduction, or any copyin Copyright (C) 1988-1992 by Cambridge University Press. Programs Copyright (C) 1988-1992 by Numerical Recipes Software. Sample page from NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING (ISBN 0-521-43108-5) g of machinereadable files (including this one) to any server computer, is strictly prohibited. To order Numerical Recipes books or CDROMs, visit website http://www.nr.com or call 1-800-872-7423 (North America only), or send email to directcustserv@cambridge.org (outside North America). 14.8 Savitzky-Golay Smoothing Filters In §13.5 we learned something about the construction and application of digital filters, but little guidance was given on which particular filter to use. That, of course, depends on what you want to accomplish by filtering. One obvious use for low-pass filters is to smooth noisy data. The premise of data smoothing is that one is measuring a variable that is both slowly varying and also corrupted by random noise. Then it can sometimes be useful to replace each data point by some kind of local average of surrounding data points. Since nearby points measure very nearly the same underlying value, averaging can reduce the level of noise without (much) biasing the value obtained. We must comment editorially that the smoothing of data lies in a murky area, beyond the fringe of some better posed, and therefore more highly recommended, techniques that are discussed elsewhere in this book. If you are fitting data to a parametric model, for example (see Chapter 15), it is almost always better to use raw data than to use data that has been pre-processed by a smoothing procedure. Another alternative to blind smoothing is so-called “optimal” or Wiener filtering, as discussed in §13.3 and more generally in §13.6. Data smoothing is probably most justified when it is used simply as a graphical technique, to guide the eye through a forest of data points all with large error bars; or as a means of making initial rough estimates of simple parameters from a graph. In this section we discuss a particular type of low-pass filter, well-adapted for data smoothing, and termed variously Savitzky-Golay [1], least-squares [2], or DISPO (Digital Smoothing Polynomial) [3] filters. Rather than having their properties defined in the Fourier domain, and then translated to the time domain, Savitzky-Golay filters derive directly from a particular formulation of the data smoothing problem in the time domain, as we will now see. Savitzky-Golay filters were initially (and are still often) used to render visible the relative widths and heights of spectral lines in noisy spectrometric data. Recall that a digital filter is applied to a series of equally spaced data values fi ≡ f(ti), where ti ≡ t0 + i∆ for some constant sample spacing ∆ and i = ... − 2, −1, 0, 1, 2,... . We have seen (§13.5) that the simplest type of digital filter (the nonrecursive or finite impulse response filter) replaces each data value fi by a linear combination gi of itself and some number of nearby neighbors, gi = nR n=−nL cnfi+n (14.8.1) Here nL is the number of points used “to the left” of a data point i, i.e., earlier than it, while nR is the number used to the right, i.e., later. A so-called causal filter would have nR = 0. As a starting point for understanding Savitzky-Golay filters, consider the simplest possible averaging procedure: For some fixed nL = nR, compute each gi as the average of the data points from fi−nL to fi+nR . This is sometimes called moving window averaging and corresponds to equation (14.8.1) with constant cn = 1/(nL + nR + 1). If the underlying function is constant, or is changing linearly with time (increasing or decreasing), then no bias is introduced into the result. Higher points at one end of the averaging interval are on the average balanced by lower points at the other end. A bias is introduced, however, if the underlying function has a nonzero second derivative. At a local maximum, for example, moving window averaging always reduces the function value. In the spectrometric application, a narrow spectral line has its height reduced and its width increased. Since these parameters are themselves of physical interest, the bias introduced is distinctly undesirable. Note, however, that moving window averaging does preserve the area under a spectral line, which is its zeroth moment, and also (if the window is symmetric with nL = nR) its mean position in time, which is its first moment. What is violated is the second moment, equivalent to the line width. The idea of Savitzky-Golay filtering is to find filter coefficients cn that preserve higher moments. Equivalently, the idea is to approximate the underlying function within the moving window not by a constant (whose estimate is the average), but by a polynomial of higher order, typically quadratic or quartic: For each point fi, we least-squares fit a polynomial to all
