Re: [R] Fitting an Inverse Gamma Distribution
Hello David, Thank you for pointing me to the GeneralizedHyperbolic package. I've been playing around with hyperbFit with my data, but get: Warning Messages: In besselK(zeta, nu = 1) : value out of range in 'bessel_k' 10 Next, just to see if I could get something to plot I tried the following to no avail: hist(iniSal_US_forHist,breaks=seq(1.1,21,by=0.625),col=grey,freq=F,xlim=c(0,21)) curve(dgig(x,param=c(11,0,-4)),col=blue,from=0,to=20,add=T) Could I trouble you for a bit more guidance? Also, does the fitting function also deal with shifts? In other words, if 500 values were sampled from an inv. gamma and and then had 4 added to them, would the fitting function account for this? Thanks, Eric -- View this message in context: http://r.789695.n4.nabble.com/Fitting-an-Inverse-Gamma-Distribution-tp3216865p3218474.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Fitting an Inverse Gamma Distribution
http://r.789695.n4.nabble.com/file/n3216865/Inverse_Gamma.png Hello, I am seeking help in estimating the parameters of an inverse gamma distribution (from the 'actuar' package) using a function like 'fitdistr'. Unfortunately I haven't found such a package using findFn('fit Inverse Gamma') from the 'sos' package and was therefore hoping someone might be aware of such a function? Secondly, is there a way to shift the pdf (code below) to the right (rather than the data to the left)? I tried: par(new=T) shift-1 hist(iniSal_US_forHist,breaks=seq(1.1,21,by=0.625),col=grey,freq=F,xlim=c(0-shift,21-shift),plot=F) curve(dinvgamma(x,scale=11.835,shape=4.4242),from=0,to=20,add=T,col=purple,lwd=2) but this failed in shifting the curve. More broadly, the data plotted in the histogram represents a calibration target for output generated by a finite difference solute transport model. The values that will be generated by the finite difference model would ideally fit the shifted (if that can be figured out) inverse gamma pdf. To the extent that fit is deemed poor, the parameter estimation software associated with the finite difference model will adjust parameters until a better fit is found. I will try to use a goodness of fit test to determine if one set of parameter values in the finite difference model produce output that fits the inv. gamma curve herein better than another. But first I need to establish the curve the finite difference model output should target and would greatly appreciate any leads on how it might be shifted and/or fit more precisely by a fitdistr-like function. library(MASS) library(actuar) iniSal_US_forHist-c(2.368000,3.532614,3.064330,3.347069,3.066333,4.233636,3.465650,2.858553, 2.946731,2.945417,2.415000,2.873019,5.521000,5.788148,5.314630,5.509672,6.032840,6.009310, 4.110833,6.073182,5.652833,4.425733,6.481852,4.076857,3.289310,4.524000,3.985811,5.399714, 4.490606,6.956729,5.270933,8.099107,5.058250,6.394500,5.644000,5.202459,5.67,3.152680, 3.220952,2.777381,3.115467,3.642759,3.488333,3.022439,2.610290,2.618571,3.218000,3.417634, 10.327317,7.344270,6.886154,4.015800,3.063103,6.832292,4.600238,2.939000,5.999027,7.894878, 4.411538,2.384762,6.816154,2.782500,2.475333,2.799138,2.739063,2.619917,2.892545,2.468167, 2.577079,2.821875,2.502500,2.969032,2.046023,3.073077,4.408000,3.411774,3.50,4.283607, 4.284000,4.276714,3.228103,2.639875,3.453194,2.821200,3.838723,1.714253,2.273750,2.611882, 2.321781,2.567500,2.557045,1.288875,2.175211,1.736000,2.250781,7.433366,7.033553,5.47, 7.132727,8.505937,9.174545,6.554487,7.060286,6.617160,8.210986,4.404045,6.062381,5.149625, 2.972105,5.358889,3.910968,3.715873,1.728966,2.843667,4.413906,3.016346,7.168636,3.839394, 3.930141,7.019882,3.459429,5.050250,3.492714,3.226667,3.987667,2.770227,3.661167,1.553000, 2.867391,2.897193,2.611707,2.577167,2.904697,2.733077,2.507241,11.044865,6.425484,8.567222, 8.552344,7.493396,4.807381,9.697869,9.471333,6.783175,4.563571,8.059649,9.448679,5.803778, 4.769423,4.424634,7.586042,4.451556,3.622373,6.390152,4.424375,4.135806,5.025400,5.410635, 7.012292,2.961071,3.192188,2.989643,3.471429,2.867966,1.980541,3.172344,2.574783,2.958983, 1.708140,3.604853,3.479000,2.845000,2.742603,2.923968,3.620308,2.452500,2.721375,3.166333, 2.742162,2.793000,3.337000,5.192025,5.365875,3.079000,8.415970,6.612277,6.734706,4.856857, 5.164783,7.743667,6.894151,4.666538,9.227167,8.077581,6.109833,6.621724,18.098182,12.705600, 15.490784,17.394750,12.422364,14.832727,8.326000,11.352400,3.431429,2.658261,3.219773,3.605185, 4.030299,3.262241,3.503250,3.522763,2.847312,2.996618,3.075769,3.387731,3.066923,3.078200, 2.466957,3.214167,2.707778,3.384839,2.283556,2.912258,3.378000,2.726750,2.95,2.195000, 4.819063,3.604578,3.694906,5.068000,4.676582,3.028831,4.261042,3.593235,4.501224,2.880317, 5.750333,3.257833,3.967458,2.522292,2.725738,2.549231,2.591389,2.990488,2.681222,2.685854, 2.284750,2.585938,2.432824,3.108875,2.611340,3.916667,2.418095,2.476406,2.801235,3.278000, 2.434921,2.617826,3.133939,2.774321,4.196173,3.764286,3.555833,5.317361,3.970800,4.136400, 4.487013,3.746393,4.754000,3.854316,3.742353,3.044079,2.817821,3.995179,3.643134,3.642593, 3.604533,2.935902,4.088310,5.344407,3.076883,3.287105,3.720870,2.032258,2.872593,5.787313, 6.017838,5.425205,4.880600,3.582295,4.90,3.489016,4.603030,5.344407,6.184286,4.047083, 4.788304,4.661325,4.815938,4.056790,3.765595,5.348772,5.200222,4.906311,3.900147,3.782897, 3.767313,3.417732,3.725455,2.888750,2.552333,2.521613,2.531522,2.510833,2.710208,2.445273, 2.619750,2.094737,2.399355,2.758000,2.317077,2.247755,3.594333,4.607805,2.69,3.084706, 3.529000,2.326200,3.309851,2.647805,2.802250,2.778667,3.231235,2.418065,3.134545,3.807843, 2.988372,2.709792,3.084035,3.633279,2.958750,2.170081,2.67,2.640706,2.841600,3.399219, 2.561373,2.574824,3.046447,2.647500,3.554875,1.865000,2.858333,2.355000,2.508082,2.376833,
Re: [R] Fitting an Inverse Gamma Distribution
On 14/01/2011 11:46 a.m., emorway wrote: http://r.789695.n4.nabble.com/file/n3216865/Inverse_Gamma.png Hello, I am seeking help in estimating the parameters of an inverse gamma distribution (from the 'actuar' package) using a function like 'fitdistr'. Unfortunately I haven't found such a package using findFn('fit Inverse Gamma') from the 'sos' package and was therefore hoping someone might be aware of such a function? In my package GeneralizedHyperbolic on R-Forge (not yet in the CRAN version) you will find a fitting routine for the generalized inverse Gaussian distribution which has the inverse Gaussian as a special case. David Scott Secondly, is there a way to shift the pdf (code below) to the right (rather than the data to the left)? I tried: par(new=T) shift-1 hist(iniSal_US_forHist,breaks=seq(1.1,21,by=0.625),col=grey,freq=F,xlim=c(0-shift,21-shift),plot=F) curve(dinvgamma(x,scale=11.835,shape=4.4242),from=0,to=20,add=T,col=purple,lwd=2) but this failed in shifting the curve. More broadly, the data plotted in the histogram represents a calibration target for output generated by a finite difference solute transport model. The values that will be generated by the finite difference model would ideally fit the shifted (if that can be figured out) inverse gamma pdf. To the extent that fit is deemed poor, the parameter estimation software associated with the finite difference model will adjust parameters until a better fit is found. I will try to use a goodness of fit test to determine if one set of parameter values in the finite difference model produce output that fits the inv. gamma curve herein better than another. But first I need to establish the curve the finite difference model output should target and would greatly appreciate any leads on how it might be shifted and/or fit more precisely by a fitdistr-like function. library(MASS) library(actuar) iniSal_US_forHist-c(2.368000,3.532614,3.064330,3.347069,3.066333,4.233636,3.465650,2.858553, 2.946731,2.945417,2.415000,2.873019,5.521000,5.788148,5.314630,5.509672,6.032840,6.009310, 4.110833,6.073182,5.652833,4.425733,6.481852,4.076857,3.289310,4.524000,3.985811,5.399714, 4.490606,6.956729,5.270933,8.099107,5.058250,6.394500,5.644000,5.202459,5.67,3.152680, 3.220952,2.777381,3.115467,3.642759,3.488333,3.022439,2.610290,2.618571,3.218000,3.417634, 10.327317,7.344270,6.886154,4.015800,3.063103,6.832292,4.600238,2.939000,5.999027,7.894878, 4.411538,2.384762,6.816154,2.782500,2.475333,2.799138,2.739063,2.619917,2.892545,2.468167, 2.577079,2.821875,2.502500,2.969032,2.046023,3.073077,4.408000,3.411774,3.50,4.283607, 4.284000,4.276714,3.228103,2.639875,3.453194,2.821200,3.838723,1.714253,2.273750,2.611882, 2.321781,2.567500,2.557045,1.288875,2.175211,1.736000,2.250781,7.433366,7.033553,5.47, 7.132727,8.505937,9.174545,6.554487,7.060286,6.617160,8.210986,4.404045,6.062381,5.149625, 2.972105,5.358889,3.910968,3.715873,1.728966,2.843667,4.413906,3.016346,7.168636,3.839394, 3.930141,7.019882,3.459429,5.050250,3.492714,3.226667,3.987667,2.770227,3.661167,1.553000, 2.867391,2.897193,2.611707,2.577167,2.904697,2.733077,2.507241,11.044865,6.425484,8.567222, 8.552344,7.493396,4.807381,9.697869,9.471333,6.783175,4.563571,8.059649,9.448679,5.803778, 4.769423,4.424634,7.586042,4.451556,3.622373,6.390152,4.424375,4.135806,5.025400,5.410635, 7.012292,2.961071,3.192188,2.989643,3.471429,2.867966,1.980541,3.172344,2.574783,2.958983, 1.708140,3.604853,3.479000,2.845000,2.742603,2.923968,3.620308,2.452500,2.721375,3.166333, 2.742162,2.793000,3.337000,5.192025,5.365875,3.079000,8.415970,6.612277,6.734706,4.856857, 5.164783,7.743667,6.894151,4.666538,9.227167,8.077581,6.109833,6.621724,18.098182,12.705600, 15.490784,17.394750,12.422364,14.832727,8.326000,11.352400,3.431429,2.658261,3.219773,3.605185, 4.030299,3.262241,3.503250,3.522763,2.847312,2.996618,3.075769,3.387731,3.066923,3.078200, 2.466957,3.214167,2.707778,3.384839,2.283556,2.912258,3.378000,2.726750,2.95,2.195000, 4.819063,3.604578,3.694906,5.068000,4.676582,3.028831,4.261042,3.593235,4.501224,2.880317, 5.750333,3.257833,3.967458,2.522292,2.725738,2.549231,2.591389,2.990488,2.681222,2.685854, 2.284750,2.585938,2.432824,3.108875,2.611340,3.916667,2.418095,2.476406,2.801235,3.278000, 2.434921,2.617826,3.133939,2.774321,4.196173,3.764286,3.555833,5.317361,3.970800,4.136400, 4.487013,3.746393,4.754000,3.854316,3.742353,3.044079,2.817821,3.995179,3.643134,3.642593, 3.604533,2.935902,4.088310,5.344407,3.076883,3.287105,3.720870,2.032258,2.872593,5.787313, 6.017838,5.425205,4.880600,3.582295,4.90,3.489016,4.603030,5.344407,6.184286,4.047083, 4.788304,4.661325,4.815938,4.056790,3.765595,5.348772,5.200222,4.906311,3.900147,3.782897, 3.767313,3.417732,3.725455,2.888750,2.552333,2.521613,2.531522,2.510833,2.710208,2.445273, 2.619750,2.094737,2.399355,2.758000,2.317077,2.247755,3.594333,4.607805,2.69,3.084706,
[R] Fitting an Inverse Gamma Distribution to Survey Data
Hello, I've been attempting to fit the data below with an inverse gamma distribution. The reason for this is outside proprietary software (@Risk) kicked back a Pearson5 (inverse gamma) as the best fitting distribution with a Chi-Sqr goodness-of-fit roughly 40% better than with a log-normal fit. Looking up Inverse gamma on this forum led me the following post: http://r.789695.n4.nabble.com/Inverse-gamma-td825481.html#a825482 But I think I'm misunderstanding the suggestion made in that post. Is there way to estimate the shape and rate parameters for an inverse-gamma and then plot the PDF as I have done below using other more readily available pdf's in R? Thanks, Eric library(MASS) iniSal_US_forHist-c(2.368000,3.532614,3.064330,3.347069,3.066333,4.233636,3.465650,2.858553, 2.946731,2.945417,2.415000,2.873019,5.521000,5.788148,5.314630,5.509672,6.032840,6.009310, 4.110833,6.073182,5.652833,4.425733,6.481852,4.076857,3.289310,4.524000,3.985811,5.399714, 4.490606,6.956729,5.270933,8.099107,5.058250,6.394500,5.644000,5.202459,5.67,3.152680, 3.220952,2.777381,3.115467,3.642759,3.488333,3.022439,2.610290,2.618571,3.218000,3.417634, 10.327317,7.344270,6.886154,4.015800,3.063103,6.832292,4.600238,2.939000,5.999027,7.894878, 4.411538,2.384762,6.816154,2.782500,2.475333,2.799138,2.739063,2.619917,2.892545,2.468167, 2.577079,2.821875,2.502500,2.969032,2.046023,3.073077,4.408000,3.411774,3.50,4.283607, 4.284000,4.276714,3.228103,2.639875,3.453194,2.821200,3.838723,1.714253,2.273750,2.611882, 2.321781,2.567500,2.557045,1.288875,2.175211,1.736000,2.250781,7.433366,7.033553,5.47, 7.132727,8.505937,9.174545,6.554487,7.060286,6.617160,8.210986,4.404045,6.062381,5.149625, 2.972105,5.358889,3.910968,3.715873,1.728966,2.843667,4.413906,3.016346,7.168636,3.839394, 3.930141,7.019882,3.459429,5.050250,3.492714,3.226667,3.987667,2.770227,3.661167,1.553000, 2.867391,2.897193,2.611707,2.577167,2.904697,2.733077,2.507241,11.044865,6.425484,8.567222, 8.552344,7.493396,4.807381,9.697869,9.471333,6.783175,4.563571,8.059649,9.448679,5.803778, 4.769423,4.424634,7.586042,4.451556,3.622373,6.390152,4.424375,4.135806,5.025400,5.410635, 7.012292,2.961071,3.192188,2.989643,3.471429,2.867966,1.980541,3.172344,2.574783,2.958983, 1.708140,3.604853,3.479000,2.845000,2.742603,2.923968,3.620308,2.452500,2.721375,3.166333, 2.742162,2.793000,3.337000,5.192025,5.365875,3.079000,8.415970,6.612277,6.734706,4.856857, 5.164783,7.743667,6.894151,4.666538,9.227167,8.077581,6.109833,6.621724,18.098182,12.705600, 15.490784,17.394750,12.422364,14.832727,8.326000,11.352400,3.431429,2.658261,3.219773,3.605185, 4.030299,3.262241,3.503250,3.522763,2.847312,2.996618,3.075769,3.387731,3.066923,3.078200, 2.466957,3.214167,2.707778,3.384839,2.283556,2.912258,3.378000,2.726750,2.95,2.195000, 4.819063,3.604578,3.694906,5.068000,4.676582,3.028831,4.261042,3.593235,4.501224,2.880317, 5.750333,3.257833,3.967458,2.522292,2.725738,2.549231,2.591389,2.990488,2.681222,2.685854, 2.284750,2.585938,2.432824,3.108875,2.611340,3.916667,2.418095,2.476406,2.801235,3.278000, 2.434921,2.617826,3.133939,2.774321,4.196173,3.764286,3.555833,5.317361,3.970800,4.136400, 4.487013,3.746393,4.754000,3.854316,3.742353,3.044079,2.817821,3.995179,3.643134,3.642593, 3.604533,2.935902,4.088310,5.344407,3.076883,3.287105,3.720870,2.032258,2.872593,5.787313, 6.017838,5.425205,4.880600,3.582295,4.90,3.489016,4.603030,5.344407,6.184286,4.047083, 4.788304,4.661325,4.815938,4.056790,3.765595,5.348772,5.200222,4.906311,3.900147,3.782897, 3.767313,3.417732,3.725455,2.888750,2.552333,2.521613,2.531522,2.510833,2.710208,2.445273, 2.619750,2.094737,2.399355,2.758000,2.317077,2.247755,3.594333,4.607805,2.69,3.084706, 3.529000,2.326200,3.309851,2.647805,2.802250,2.778667,3.231235,2.418065,3.134545,3.807843, 2.988372,2.709792,3.084035,3.633279,2.958750,2.170081,2.67,2.640706,2.841600,3.399219, 2.561373,2.574824,3.046447,2.647500,3.554875,1.865000,2.858333,2.355000,2.508082,2.376833, 2.728710,1.752833,1.571967,1.800333,2.265455,2.353226,2.568028,2.359500,2.472639,1.675385, 2.667386,2.49,2.482632,2.593452,2.695510,2.466941,2.624211,3.835645,3.519667,2.661940, 3.516167,3.146585,3.420462,4.809833,3.028500,3.192297,4.256333,2.516897,3.033846,2.793359, 6.700714,5.441250,6.872500,4.528333,7.490328,4.673788,6.376885,3.023167,4.429167,4.317625, 16.729231,8.372500,6.279828,10.805098,8.359452,8.277576,8.008846,8.742308,12.155942,5.975063, 3.317333,2.883021,3.310822,3.119219,3.921190,3.552986,3.647162,4.017667,3.895342,3.381096, 2.769412,3.225294,4.169333,3.733919,2.859492,3.674875,3.401017,3.160267,4.109545,4.347867, 3.867000,3.672763,4.364844,3.804250,2.613000,4.289909,4.212059,4.785797,4.149687,6.299444, 5.640135,5.713448,4.766438,7.032027,5.610533,3.126111,6.322029,4.417692,6.392532,2.753175, 2.549655,3.456533,3.199000,3.852338,3.387549,3.098033,3.271733,3.679180,2.655484,6.858167, 5.808033,7.55,8.388387,5.108732,7.895152,5.223580,5.741714,8.026250,4.928421,2.797292,