R Under development (unstable) (2024-02-02 r85855 ucrt) -- "Unsuffered Consequences" Copyright (C) 2024 The R Foundation for Statistical Computing Platform: x86_64-w64-mingw32/x64 R is free software and comes with ABSOLUTELY NO WARRANTY. You are welcome to redistribute it under certain conditions. Type 'license()' or 'licence()' for distribution details. R is a collaborative project with many contributors. Type 'contributors()' for more information and 'citation()' on how to cite R or R packages in publications. Type 'demo()' for some demos, 'help()' for on-line help, or 'help.start()' for an HTML browser interface to help. Type 'q()' to quit R. > #### OOps! Running this in 'CMD check' or in *R* __for the first time__ > #### ===== gives a wrong result (at the end) than when run a 2nd time > ####-- problem disappears with introduction of if (psw) call ... in Fortran > > suppressMessages(library(cobs)) > options(digits = 6) > if(!dev.interactive(orNone=TRUE)) pdf("ex1.pdf") > > source(system.file("util.R", package = "cobs")) > > ## Simple example from example(cobs) > set.seed(908) > x <- seq(-1,1, len = 50) > f.true <- pnorm(2*x) > y <- f.true + rnorm(50)/10 > ## specify constraints (boundary conditions) > con <- rbind(c( 1,min(x),0), + c(-1,max(x),1), + c( 0, 0, 0.5)) > ## obtain the median *regression* B-spline using automatically selected knots > coR <- cobs(x,y,constraint = "increase", pointwise = con) qbsks2(): Performing general knot selection ... Deleting unnecessary knots ... > summaryCobs(coR) List of 24 $ call : language cobs(x = x, y = y, constraint = "increase", pointwise = con) $ tau : num 0.5 $ degree : num 2 $ constraint : chr "increase" $ ic : chr "AIC" $ pointwise : num [1:3, 1:3] 1 -1 0 -1 1 0 0 1 0.5 $ select.knots : logi TRUE $ select.lambda: logi FALSE $ x : num [1:50] -1 -0.959 -0.918 -0.878 -0.837 ... $ y : num [1:50] 0.2254 0.0916 0.0803 -0.0272 -0.0454 ... $ resid : num [1:50] 0.1976 0.063 0.0491 -0.0626 -0.0868 ... $ fitted : num [1:50] 0.0278 0.0287 0.0312 0.0354 0.0414 ... $ coef : num [1:4] 0.0278 0.0278 0.8154 1 $ knots : num [1:3] -1 -0.224 1 $ k0 : num 4 $ k : num 4 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 6.19 $ lambda : num 0 $ icyc : int 7 $ ifl : int 1 $ pp.lambda : NULL $ pp.sic : NULL $ i.mask : NULL cb.lo ci.lo fit ci.up cb.up 1 -6.77514e-02 -0.029701622 0.0278152 0.0853320 0.123382 2 -6.41787e-02 -0.027468888 0.0280224 0.0835138 0.120224 3 -6.04433e-02 -0.024973163 0.0286442 0.0822615 0.117732 4 -5.65412e-02 -0.022212175 0.0296803 0.0815728 0.115902 5 -5.24674e-02 -0.019182756 0.0311310 0.0814447 0.114729 6 -4.82149e-02 -0.015880775 0.0329961 0.0818729 0.114207 7 -4.37751e-02 -0.012301110 0.0352757 0.0828524 0.114326 8 -3.91381e-02 -0.008437641 0.0379697 0.0843771 0.115077 9 -3.42918e-02 -0.004283290 0.0410782 0.0864397 0.116448 10 -2.92233e-02 0.000169901 0.0446012 0.0890325 0.118426 11 -2.39179e-02 0.004930665 0.0485387 0.0921467 0.120995 12 -1.83600e-02 0.010008360 0.0528906 0.0957728 0.124141 13 -1.25335e-02 0.015412811 0.0576570 0.0999012 0.127847 14 -6.42140e-03 0.021154129 0.0628378 0.1045216 0.132097 15 -6.81378e-06 0.027242531 0.0684332 0.1096238 0.136873 16 6.72715e-03 0.033688168 0.0744430 0.1151978 0.142159 17 1.37970e-02 0.040500961 0.0808672 0.1212335 0.147938 18 2.12185e-02 0.047690461 0.0877060 0.1277215 0.154193 19 2.90068e-02 0.055265726 0.0949592 0.1346527 0.160912 20 3.71760e-02 0.063235225 0.1026269 0.1420185 0.168078 21 4.57390e-02 0.071606758 0.1107090 0.1498113 0.175679 22 5.47075e-02 0.080387396 0.1192056 0.1580238 0.183704 23 6.40921e-02 0.089583438 0.1281167 0.1666500 0.192141 24 7.39018e-02 0.099200377 0.1374422 0.1756841 0.200983 25 8.41444e-02 0.109242876 0.1471823 0.1851216 0.210220 26 9.48262e-02 0.119714746 0.1573367 0.1949588 0.219847 27 1.05952e-01 0.130618921 0.1679057 0.2051925 0.229859 28 1.17526e-01 0.141957438 0.1788891 0.2158208 0.240253 29 1.29548e-01 0.153731401 0.1902870 0.2268426 0.251026 30 1.42021e-01 0.165940947 0.2020994 0.2382578 0.262178 31 1.54941e-01 0.178585191 0.2143262 0.2500672 0.273711 32 1.68306e-01 0.191662165 0.2269675 0.2622729 0.285629 33 1.82111e-01 0.205168744 0.2400233 0.2748778 0.297936 34 1.96348e-01 0.219100556 0.2534935 0.2878865 0.310639 35 2.11008e-01 0.233451886 0.2673782 0.3013046 0.323748 36 2.26079e-01 0.248215565 0.2816774 0.3151392 0.337276 37 2.41547e-01 0.263382876 0.2963910 0.3293992 0.351235 38 2.57393e-01 0.278943451 0.3115191 0.3440948 0.365645 39 2.73599e-01 0.294885220 0.3270617 0.3592382 0.380524 40 2.90023e-01 0.311080514 0.3429107 0.3747410 0.395798 41 3.06194e-01 0.327075735 0.3586411 0.3902065 0.411088 42 3.22074e-01 0.342831649 0.3742095 0.4055873 0.426345 43 3.37676e-01 0.358355597 0.3896158 0.4208761 0.441556 44 3.53012e-01 0.373655096 0.4048602 0.4360653 0.456709 45 3.68094e-01 0.388737688 0.4199426 0.4511475 0.471791 46 3.82936e-01 0.403610792 0.4348630 0.4661151 0.486790 47 3.97549e-01 0.418281590 0.4496214 0.4809611 0.501694 48 4.11944e-01 0.432756923 0.4642177 0.4956786 0.516491 49 4.26133e-01 0.447043216 0.4786521 0.5102611 0.531172 50 4.40124e-01 0.461146429 0.4929245 0.5247027 0.545725 51 4.53927e-01 0.475072016 0.5070350 0.5389979 0.560143 52 4.67551e-01 0.488824911 0.5209834 0.5531418 0.574416 53 4.81002e-01 0.502409521 0.5347698 0.5671300 0.588538 54 4.94287e-01 0.515829730 0.5483942 0.5809587 0.602501 55 5.07412e-01 0.529088909 0.5618566 0.5946243 0.616302 56 5.20381e-01 0.542189933 0.5751571 0.6081242 0.629933 57 5.33198e-01 0.555135196 0.5882955 0.6214558 0.643393 58 5.45867e-01 0.567926630 0.6012719 0.6346172 0.656677 59 5.58390e-01 0.580565721 0.6140864 0.6476070 0.669782 60 5.70769e-01 0.593053527 0.6267388 0.6604241 0.682708 61 5.83005e-01 0.605390690 0.6392293 0.6730679 0.695454 62 5.95098e-01 0.617577451 0.6515577 0.6855380 0.708017 63 6.07048e-01 0.629613656 0.6637242 0.6978347 0.720400 64 6.18854e-01 0.641498766 0.6757287 0.7099586 0.732603 65 6.30515e-01 0.653231865 0.6875711 0.7219104 0.744627 66 6.42028e-01 0.664811658 0.6992516 0.7336916 0.756475 67 6.53391e-01 0.676236478 0.7107701 0.7453037 0.768149 68 6.64600e-01 0.687504287 0.7221266 0.7567489 0.779653 69 6.75652e-01 0.698612675 0.7333211 0.7680295 0.790991 70 6.86541e-01 0.709558867 0.7443536 0.7791483 0.802166 71 6.97262e-01 0.720339721 0.7552241 0.7901084 0.813186 72 7.07810e-01 0.730951740 0.7659326 0.8009134 0.824055 73 7.18179e-01 0.741391078 0.7764791 0.8115671 0.834779 74 7.28361e-01 0.751653555 0.7868636 0.8220736 0.845367 75 7.38348e-01 0.761734678 0.7970861 0.8324375 0.855824 76 7.48134e-01 0.771629669 0.8071466 0.8426636 0.866160 77 7.57709e-01 0.781333498 0.8170452 0.8527568 0.876382 78 7.67065e-01 0.790840929 0.8267817 0.8627224 0.886499 79 7.76192e-01 0.800146569 0.8363562 0.8725659 0.896520 80 7.85083e-01 0.809244928 0.8457688 0.8822926 0.906455 81 7.93727e-01 0.818130488 0.8550193 0.8919081 0.916312 82 8.02116e-01 0.826797774 0.8641079 0.9014179 0.926100 83 8.10240e-01 0.835241429 0.8730344 0.9108274 0.935829 84 8.18091e-01 0.843456291 0.8817990 0.9201417 0.945507 85 8.25661e-01 0.851437463 0.8904015 0.9293656 0.955142 86 8.32942e-01 0.859180385 0.8988421 0.9385038 0.964742 87 8.39928e-01 0.866680887 0.9071207 0.9475605 0.974313 88 8.46612e-01 0.873935236 0.9152373 0.9565393 0.983862 89 8.52989e-01 0.880940170 0.9231918 0.9654435 0.993395 90 8.59054e-01 0.887692913 0.9309844 0.9742760 1.002915 91 8.64803e-01 0.894191180 0.9386150 0.9830389 1.012427 92 8.70233e-01 0.900433167 0.9460836 0.9917341 1.021934 93 8.75343e-01 0.906417527 0.9533902 1.0003629 1.031437 94 8.80130e-01 0.912143340 0.9605348 1.0089263 1.040939 95 8.84594e-01 0.917610075 0.9675174 1.0174248 1.050441 96 8.88735e-01 0.922817542 0.9743381 1.0258586 1.059942 97 8.92551e-01 0.927765853 0.9809967 1.0342275 1.069442 98 8.96045e-01 0.932455371 0.9874933 1.0425312 1.078941 99 8.99218e-01 0.936886669 0.9938279 1.0507692 1.088438 100 9.02069e-01 0.941060487 1.0000006 1.0589406 1.097932 knots : [1] -1.00000 -0.22449 1.00000 coef : [1] 0.0278152 0.0278152 0.8153868 1.0000006 > coR1 <- cobs(x,y,constraint = "increase", pointwise = con, degree = 1) qbsks2(): Performing general knot selection ... Deleting unnecessary knots ... > summary(coR1) COBS regression spline (degree = 1) from call: cobs(x = x, y = y, constraint = "increase", degree = 1, pointwise = con) {tau=0.5}-quantile; dimensionality of fit: 4 from {4} x$knots[1:4]: -1.000002, -0.632653, 0.183673, 1.000002 with 3 pointwise constraints coef[1:4]: 0.0504467, 0.0504467, 0.6305155, 1.0000009 R^2 = 93.83% ; empirical tau (over all): 21/50 = 0.42 (target tau= 0.5) > > ## compute the median *smoothing* B-spline using automatically chosen lambda > coS <- cobs(x,y,constraint = "increase", pointwise = con, + lambda = -1, trace = 3) Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. loo.design2(): -> Xeq 51 x 22 (nz = 151 =^= 0.13%) Xieq 62 x 22 (nz = 224 =^= 0.16%) ........................ The algorithm has converged. You might plot() the returned object (which plots 'sic' against 'lambda') to see if you have found the global minimum of the information criterion so that you can determine if you need to adjust any or all of 'lambda.lo', 'lambda.hi' and 'lambda.length' and refit the model. > with(coS, cbind(pp.lambda, pp.sic, k0, ifl, icyc)) pp.lambda pp.sic k0 ifl icyc [1,] 3.54019e-05 -2.64644 22 1 21 [2,] 6.92936e-05 -2.64644 22 1 21 [3,] 1.35631e-04 -2.64644 22 1 20 [4,] 2.65477e-04 -2.64644 22 1 22 [5,] 5.19629e-04 -2.64644 22 1 22 [6,] 1.01709e-03 -2.64644 22 1 23 [7,] 1.99080e-03 -2.68274 21 1 20 [8,] 3.89667e-03 -2.75212 19 1 18 [9,] 7.62711e-03 -2.73932 19 1 14 [10,] 1.49289e-02 -2.85261 16 1 13 [11,] 2.92209e-02 -2.97873 12 1 12 [12,] 5.71953e-02 -3.01058 11 1 12 [13,] 1.11951e-01 -3.04364 10 1 11 [14,] 2.19126e-01 -3.11242 8 1 12 [15,] 4.28904e-01 -3.17913 6 1 12 [16,] 8.39512e-01 -3.18824 5 1 11 [17,] 1.64321e+00 -3.01467 5 1 12 [18,] 3.21633e+00 -3.01380 4 1 11 [19,] 6.29545e+00 -3.01380 4 1 10 [20,] 1.23223e+01 -3.01380 4 1 11 [21,] 2.41190e+01 -3.01380 4 1 11 [22,] 4.72092e+01 -3.01380 4 1 10 [23,] 9.24046e+01 -3.01380 4 1 10 [24,] 1.80867e+02 -3.01380 4 1 10 [25,] 3.54019e+02 -3.01380 4 1 10 > with(coS, plot(pp.sic ~ pp.lambda, type = "b", log = "x", col=2, + main = deparse(call))) > ##-> very nice minimum close to 1 > > summaryCobs(coS) List of 24 $ call : language cobs(x = x, y = y, constraint = "increase", lambda = -1, pointwise = con, trace = 3) $ tau : num 0.5 $ degree : num 2 $ constraint : chr "increase" $ ic : NULL $ pointwise : num [1:3, 1:3] 1 -1 0 -1 1 0 0 1 0.5 $ select.knots : logi TRUE $ select.lambda: logi TRUE $ x : num [1:50] -1 -0.959 -0.918 -0.878 -0.837 ... $ y : num [1:50] 0.2254 0.0916 0.0803 -0.0272 -0.0454 ... $ resid : num [1:50] 0.2254 0.0829 0.062 -0.0562 -0.0862 ... $ fitted : num [1:50] 0 0.00869 0.01837 0.02906 0.04075 ... $ coef : num [1:22] 0 0.00819 0.03365 0.06662 0.10458 ... $ knots : num [1:20] -1 -0.918 -0.796 -0.714 -0.592 ... $ k0 : int [1:25] 22 22 22 22 22 22 21 19 19 16 ... $ k : int 5 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 6.19 $ lambda : Named num 0.84 ..- attr(*, "names")= chr "lambda" $ icyc : int [1:25] 21 21 20 22 22 23 20 18 14 13 ... $ ifl : int [1:25] 1 1 1 1 1 1 1 1 1 1 ... $ pp.lambda : num [1:25] 0 0 0 0 0.001 0.001 0.002 0.004 0.008 0.015 ... $ pp.sic : num [1:25] -2.65 -2.65 -2.65 -2.65 -2.65 ... $ i.mask : logi [1:25] TRUE TRUE TRUE TRUE TRUE TRUE ... cb.lo ci.lo fit ci.up cb.up 1 -0.07071332 -0.03907635 -3.77249e-07 0.0390756 0.0707126 2 -0.06555125 -0.03435600 4.17438e-03 0.0427048 0.0739000 3 -0.06016465 -0.02940203 8.59400e-03 0.0465900 0.0773526 4 -0.05455349 -0.02421442 1.32585e-02 0.0507314 0.0810704 5 -0.04871809 -0.01879334 1.81678e-02 0.0551289 0.0850537 6 -0.04265897 -0.01313909 2.33220e-02 0.0597831 0.0893029 7 -0.03637554 -0.00725134 2.87210e-02 0.0646934 0.0938176 8 -0.02986704 -0.00112966 3.43649e-02 0.0698595 0.0985969 9 -0.02313305 0.00522618 4.02537e-02 0.0752812 0.1036404 10 -0.01617351 0.01181620 4.63873e-02 0.0809584 0.1089481 11 -0.00898880 0.01864020 5.27658e-02 0.0868914 0.1145204 12 -0.00157983 0.02569768 5.93891e-02 0.0930806 0.1203581 13 0.00605308 0.03298846 6.62573e-02 0.0995262 0.1264615 14 0.01391000 0.04051257 7.33704e-02 0.1062282 0.1328307 15 0.02199057 0.04826981 8.07283e-02 0.1131867 0.1394660 16 0.03029461 0.05626010 8.83310e-02 0.1204020 0.1463675 17 0.03882336 0.06448412 9.61787e-02 0.1278732 0.1535339 18 0.04757769 0.07294234 1.04271e-01 0.1355999 0.1609646 19 0.05655804 0.08163500 1.12608e-01 0.1435819 0.1686589 20 0.06576441 0.09056212 1.21191e-01 0.1518192 0.1766169 21 0.07519637 0.09972344 1.30018e-01 0.1603120 0.1848391 22 0.08485262 0.10911826 1.39090e-01 0.1690610 0.1933266 23 0.09473211 0.11874598 1.48406e-01 0.1780668 0.2020807 24 0.10483493 0.12860668 1.57968e-01 0.1873294 0.2111011 25 0.11516076 0.13870015 1.67775e-01 0.1968489 0.2203882 26 0.12570956 0.14902638 1.77826e-01 0.2066253 0.2299421 27 0.13648327 0.15958645 1.88122e-01 0.2166576 0.2397608 28 0.14748286 0.17038090 1.98663e-01 0.2269453 0.2498433 29 0.15870881 0.18140998 2.09449e-01 0.2374880 0.2601892 30 0.17016110 0.19267368 2.20480e-01 0.2482859 0.2707984 31 0.18183922 0.20417172 2.31755e-01 0.2593391 0.2816716 32 0.19374227 0.21590361 2.43276e-01 0.2706482 0.2928095 33 0.20587062 0.22786955 2.55041e-01 0.2822129 0.3042118 34 0.21822524 0.24007008 2.67051e-01 0.2940328 0.3158776 35 0.23080666 0.25250549 2.79306e-01 0.3061075 0.3278063 36 0.24361488 0.26517577 2.91806e-01 0.3184370 0.3399979 37 0.25664938 0.27808064 3.04551e-01 0.3310217 0.3524530 38 0.26990862 0.29121926 3.17541e-01 0.3438624 0.3651730 39 0.28339034 0.30459037 3.30775e-01 0.3569602 0.3781603 40 0.29709467 0.31819405 3.44255e-01 0.3703152 0.3914146 41 0.31102144 0.33203019 3.57979e-01 0.3839275 0.4049363 42 0.32517059 0.34609876 3.71948e-01 0.3977971 0.4187252 43 0.33954481 0.36040126 3.86162e-01 0.4119224 0.4327789 44 0.35414537 0.37493839 4.00621e-01 0.4263028 0.4470958 45 0.36897279 0.38971043 4.15324e-01 0.4409381 0.4616757 46 0.38402708 0.40471738 4.30273e-01 0.4558281 0.4765184 47 0.39930767 0.41995895 4.45466e-01 0.4709732 0.4916245 48 0.41479557 0.43541678 4.60887e-01 0.4863568 0.5069780 49 0.43039487 0.45099622 4.76442e-01 0.5018872 0.5224885 50 0.44609197 0.46668362 4.92117e-01 0.5175506 0.5381422 51 0.46188684 0.48247895 5.07913e-01 0.5333471 0.5539392 52 0.47773555 0.49833835 5.23786e-01 0.5492329 0.5698357 53 0.49336687 0.51398935 5.39461e-01 0.5649325 0.5855550 54 0.50873469 0.52938518 5.54891e-01 0.5803975 0.6010480 55 0.52383955 0.54452615 5.70077e-01 0.5956277 0.6163143 56 0.53868141 0.55941225 5.85018e-01 0.6106231 0.6313539 57 0.55325974 0.57404316 5.99714e-01 0.6253839 0.6461673 58 0.56757320 0.58841816 6.14165e-01 0.6399109 0.6607558 59 0.58161907 0.60253574 6.28371e-01 0.6542056 0.6751223 60 0.59539741 0.61639593 6.42332e-01 0.6682680 0.6892665 61 0.60890835 0.62999881 6.56048e-01 0.6820980 0.7031884 62 0.62215175 0.64334429 6.69520e-01 0.6956957 0.7168882 63 0.63512996 0.65643368 6.82747e-01 0.7090597 0.7303634 64 0.64784450 0.66926783 6.95729e-01 0.7221893 0.7436126 65 0.66029589 0.68184700 7.08466e-01 0.7350841 0.7566352 66 0.67248408 0.69417118 7.20958e-01 0.7477442 0.7694313 67 0.68440855 0.70624008 7.33205e-01 0.7601699 0.7820014 68 0.69606829 0.71805313 7.45207e-01 0.7723617 0.7943465 69 0.70746295 0.72961016 7.56965e-01 0.7843198 0.8064670 70 0.71859343 0.74091165 7.68478e-01 0.7960438 0.8183620 71 0.72946023 0.75195789 7.79746e-01 0.8075332 0.8300309 72 0.74006337 0.76274887 7.90769e-01 0.8187883 0.8414738 73 0.75040233 0.77328433 8.01547e-01 0.8298091 0.8526911 74 0.76047612 0.78356369 8.12080e-01 0.8405963 0.8636839 75 0.77028266 0.79358583 8.22368e-01 0.8511510 0.8744542 76 0.77982200 0.80335076 8.32412e-01 0.8614732 0.8850020 77 0.78909446 0.81285866 8.42211e-01 0.8715627 0.8953269 78 0.79809990 0.82210946 8.51765e-01 0.8814196 0.9054292 79 0.80683951 0.83110382 8.61074e-01 0.8910433 0.9153076 80 0.81531459 0.83984244 8.70138e-01 0.9004329 0.9249608 81 0.82352559 0.84832559 8.78957e-01 0.9095884 0.9343884 82 0.83147249 0.85655324 8.87531e-01 0.9185095 0.9435903 83 0.83915483 0.86452515 8.95861e-01 0.9271968 0.9525671 84 0.84657171 0.87224082 9.03946e-01 0.9356505 0.9613196 85 0.85372180 0.87969951 9.11786e-01 0.9438715 0.9698492 86 0.86060525 0.88690131 9.19381e-01 0.9518597 0.9781558 87 0.86722242 0.89384640 9.26731e-01 0.9596149 0.9862389 88 0.87357322 0.90053476 9.33836e-01 0.9671371 0.9940986 89 0.87965804 0.90696658 9.40696e-01 0.9744261 1.0017347 90 0.88547781 0.91314239 9.47312e-01 0.9814814 1.0091460 91 0.89103290 0.91906239 9.53683e-01 0.9883028 1.0163323 92 0.89632328 0.92472655 9.59808e-01 0.9948904 1.0232937 93 0.90134850 0.93013464 9.65689e-01 1.0012443 1.0300304 94 0.90610776 0.93528622 9.71326e-01 1.0073650 1.0365434 95 0.91060065 0.94018104 9.76717e-01 1.0132527 1.0428331 96 0.91482784 0.94481950 9.81863e-01 1.0189071 1.0488987 97 0.91878971 0.94920179 9.86765e-01 1.0243279 1.0547400 98 0.92248624 0.95332789 9.91422e-01 1.0295152 1.0603569 99 0.92591703 0.95719761 9.95833e-01 1.0344692 1.0657498 100 0.92908136 0.96081053 1.00000e+00 1.0391902 1.0709194 knots : [1] -1.0000020 -0.9183673 -0.7959184 -0.7142857 -0.5918367 -0.5102041 [7] -0.3877551 -0.2653061 -0.1836735 -0.0612245 0.0204082 0.1428571 [13] 0.2244898 0.3469388 0.4693878 0.5510204 0.6734694 0.7551020 [19] 0.8775510 1.0000020 coef : [1] -4.01161e-07 8.18714e-03 3.36534e-02 6.66159e-02 1.04576e-01 [6] 1.50032e-01 2.00486e-01 2.70027e-01 3.35473e-01 4.05918e-01 [11] 4.83858e-01 5.64259e-01 6.37163e-01 7.05069e-01 7.77561e-01 [16] 8.30474e-01 8.78390e-01 9.18810e-01 9.54232e-01 9.87743e-01 [21] 1.00000e+00 5.99960e-01 > > plot(x, y, main = "cobs(x,y, constraint=\"increase\", pointwise = *)") > matlines(x, cbind(fitted(coR), fitted(coR1), fitted(coS)), + col = 2:4, lty=1) > > ##-- real data example (still n = 50) > data(cars) > attach(cars) > co1 <- cobs(speed, dist, "increase") qbsks2(): Performing general knot selection ... Deleting unnecessary knots ... > co1.1 <- cobs(speed, dist, "increase", knots.add = TRUE) qbsks2(): Performing general knot selection ... Deleting unnecessary knots ... Searching for missing knots ... > co1.2 <- cobs(speed, dist, "increase", knots.add = TRUE, repeat.delete.add = TRUE) qbsks2(): Performing general knot selection ... Deleting unnecessary knots ... Searching for missing knots ... > ## These three all give the same -- only remaining knots (outermost data): > ic <- which("call" == names(co1)) > stopifnot(all.equal(co1[-ic], co1.1[-ic]), + all.equal(co1[-ic], co1.2[-ic])) > 1 - sum(co1 $ resid ^2) / sum((dist - mean(dist))^2) # R^2 = 64.2% [1] 0.642288 > > co2 <- cobs(speed, dist, "increase", lambda = -1)# 6 warnings Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. WARNING: Some lambdas had problems in rq.fit.sfnc(): lambda icyc ifl fidel sum|res|_s k [1,] 2.30776 16 18 250.3 7.5999 11 The algorithm has converged. You might plot() the returned object (which plots 'sic' against 'lambda') to see if you have found the global minimum of the information criterion so that you can determine if you need to adjust any or all of 'lambda.lo', 'lambda.hi' and 'lambda.length' and refit the model. Warning message: In cobs(speed, dist, "increase", lambda = -1) : drqssbc2(): Not all flags are normal (== 1), ifl : 11111111111811111111111111 > summaryCobs(co2) List of 24 $ call : language cobs(x = speed, y = dist, constraint = "increase", lambda = -1) $ tau : num 0.5 $ degree : num 2 $ constraint : chr "increase" $ ic : NULL $ pointwise : NULL $ select.knots : logi TRUE $ select.lambda: logi TRUE $ x : num [1:50] 4 4 7 7 8 9 10 10 10 11 ... $ y : num [1:50] 2 10 4 22 16 10 18 26 34 17 ... $ resid : num [1:50] -4.86 3.14 -9.75 8.25 0 ... $ fitted : num [1:50] 6.86 6.86 13.75 13.75 16 ... $ coef : num [1:20] 6.86 10.37 14.88 17.12 19.55 ... $ knots : num [1:18] 4 7 8 9 10 ... $ k0 : int [1:25] 16 16 16 16 16 16 15 15 14 12 ... $ k : int 3 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 32539 $ lambda : Named num 66.3 ..- attr(*, "names")= chr "lambda" $ icyc : int [1:25] 17 17 15 16 16 16 18 16 17 19 ... $ ifl : int [1:25] 1 1 1 1 1 1 1 1 1 1 ... $ pp.lambda : num [1:25] 0 0.01 0.01 0.02 0.04 0.08 0.16 0.31 0.6 1.18 ... $ pp.sic : num [1:25] 2.23 2.23 2.23 2.23 2.23 ... $ i.mask : logi [1:25] TRUE TRUE TRUE TRUE TRUE TRUE ... cb.lo ci.lo fit ci.up cb.up 1 -18.0106902 -9.829675 6.86289 23.5554 31.7365 2 -15.9869427 -8.308682 7.35806 23.0248 30.7031 3 -14.7253903 -7.299595 7.85201 23.0036 30.4294 4 -14.0304377 -6.671152 8.34475 23.3607 30.7199 5 -13.6842238 -6.277147 8.83627 23.9497 31.3568 6 -13.4881973 -5.984332 9.32657 24.6375 32.1413 7 -13.2846859 -5.686895 9.81565 25.3182 32.9160 8 -12.9604500 -5.308840 10.30352 25.9159 33.5675 9 -12.4418513 -4.800750 10.79017 26.3811 34.0222 10 -11.6889866 -4.135844 11.27560 26.6871 34.2402 11 -10.6923973 -3.307777 11.75982 26.8274 34.2120 12 -9.4739641 -2.331232 12.24282 26.8169 33.9596 13 -8.0930597 -1.246053 12.72459 26.6952 33.5422 14 -6.6587120 -0.125409 13.20516 26.5357 33.0690 15 -5.3461743 0.913089 13.68450 26.4559 32.7152 16 -4.3525007 1.737247 14.16278 26.5883 32.6781 17 -3.5579640 2.427497 14.64024 26.8530 32.8385 18 -2.7821201 3.104937 15.11690 27.1289 33.0159 19 -1.9790693 3.800369 15.59275 27.3851 33.1646 20 -1.2507247 4.445431 16.06788 27.6903 33.3865 21 -0.6706167 4.992698 16.54814 28.1036 33.7669 22 -0.0321786 5.581929 17.03697 28.4920 34.1061 23 0.7878208 6.295824 17.53436 28.7729 34.2809 24 1.7680338 7.120056 18.04033 28.9606 34.3126 25 2.7272662 7.933027 18.55487 29.1767 34.3825 26 3.5589282 8.663206 19.07798 29.4928 34.5970 27 4.4415126 9.430376 19.60966 29.7890 34.7778 28 5.4482980 10.283717 20.14992 30.0161 34.8515 29 6.4928117 11.165196 20.69874 30.2323 34.9047 30 7.3396986 11.916867 21.25613 30.5954 35.1726 31 8.0441586 12.575775 21.82210 31.0684 35.6000 32 8.8220660 13.286792 22.39663 31.5065 35.9712 33 9.7293382 14.087444 22.97973 31.8720 36.2301 34 10.6439776 14.895860 23.57141 32.2470 36.4988 35 11.3712676 15.581364 24.17165 32.7619 36.9720 36 12.0903660 16.264191 24.78047 33.2968 37.4706 37 12.9621618 17.052310 25.39786 33.7434 37.8336 38 13.9690548 17.933912 26.02382 34.1137 38.0786 39 14.8961150 18.764757 26.65834 34.5519 38.4206 40 15.5880299 19.440617 27.30144 35.1623 39.0149 41 16.2838155 20.121892 27.95311 35.7843 39.6224 42 17.1058166 20.890689 28.61335 36.3360 40.1209 43 17.9817714 21.698514 29.28216 36.8658 40.5826 44 18.6610233 22.377150 29.95954 37.5419 41.2581 45 19.1808813 22.951637 30.64550 38.3394 42.1101 46 19.7841459 23.584917 31.34002 39.0951 42.8959 47 20.5213873 24.310926 32.04311 39.7753 43.5648 48 21.2465778 25.031668 32.75477 40.4779 44.2630 49 21.7264172 25.590574 33.47501 41.3594 45.2236 50 22.1476742 26.112985 34.20381 42.2946 46.2600 51 22.6982036 26.724968 34.94119 43.1574 47.1842 52 23.3692423 27.420644 35.68713 43.9536 48.0050 53 23.9709413 28.072605 36.44165 44.8107 48.9124 54 24.3957772 28.608693 37.20474 45.8008 50.0137 55 24.9012324 29.201704 37.97639 46.7511 51.0516 56 25.6136292 29.936410 38.75662 47.5768 51.8996 57 26.4819493 30.778576 39.54542 48.3123 52.6089 58 27.2901515 31.583215 40.34279 49.1024 53.3954 59 28.0053951 32.328289 41.14873 49.9692 54.2921 60 28.8530207 33.165023 41.96324 50.7615 55.0735 61 29.9067993 34.142924 42.78632 51.4297 55.6658 62 31.0646746 35.193503 43.61797 52.0424 56.1713 63 32.0654071 36.141443 44.45820 52.7749 56.8510 64 32.9512818 37.015121 45.30699 53.5989 57.6627 65 33.9291102 37.953328 46.16435 54.3754 58.3996 66 35.0297017 38.976740 47.03029 55.0838 59.0309 67 36.0927814 39.977797 47.90479 55.8318 59.7168 68 36.9113073 40.817553 48.78787 56.7582 60.6644 69 37.7036248 41.642540 49.67951 57.7165 61.6554 70 38.6422928 42.568561 50.57973 58.5909 62.5172 71 39.7057083 43.581119 51.48852 59.3959 63.2713 72 40.6774154 44.534950 52.40587 60.2768 64.1343 73 41.4311354 45.345310 53.33180 61.3183 65.2325 74 42.1677718 46.147024 54.26630 62.3856 66.3648 75 42.9301723 46.968847 55.20937 63.4499 67.4886 76 43.5601364 47.704612 56.16101 64.6174 68.7619 77 43.7706750 48.161720 57.12122 66.0807 70.4718 78 43.6707129 48.413272 58.09000 67.7667 72.5093 79 43.5686662 48.666244 59.06735 69.4685 74.5660 80 43.6408522 49.038961 60.05327 71.0676 76.4657 81 43.9707369 49.587438 61.04777 72.5081 78.1248 82 44.5808788 50.326813 62.05083 73.7748 79.5208 83 45.4473581 51.241035 63.06246 74.8839 80.6776 84 46.5017521 52.284184 64.08267 75.8812 81.6636 85 47.6254964 53.376692 65.11144 76.8462 82.5974 86 48.6454643 54.402376 66.14879 77.8952 83.6521 87 49.6079288 55.392288 67.19471 78.9971 84.7815 88 50.7825571 56.527401 68.24919 79.9710 85.7158 89 52.2754967 57.878951 69.31225 80.7456 86.3490 90 54.0486439 59.421366 70.38388 81.3464 86.7191 91 55.8745252 61.001989 71.46408 81.9262 87.0536 92 57.4111269 62.391297 72.55285 82.7144 87.6946 93 58.7265102 63.634965 73.65019 83.6654 88.5739 94 59.8812030 64.773613 74.75610 84.7386 89.6310 95 60.8442106 65.786441 75.87058 85.9547 90.8969 96 61.4961859 66.593355 76.99363 87.3939 92.4911 97 61.6555082 67.072471 78.12525 89.1780 94.5950 98 61.1305173 67.095165 79.26545 91.4357 97.4004 99 59.7777137 66.565137 80.41421 94.2633 101.0507 100 57.5292654 65.436863 81.57154 97.7062 105.6138 knots : [1] 3.99998 7.00000 8.00000 9.00000 10.00000 11.00000 12.00000 13.00000 [9] 14.00000 15.00000 16.00000 17.00000 18.00000 19.00000 20.00000 22.00000 [17] 23.00000 25.00002 coef : [1] 6.862887 10.368778 14.880952 17.119048 19.547619 22.166667 24.976190 [8] 27.976190 31.166667 34.547619 38.119048 41.880952 45.833333 49.976190 [15] 54.309524 61.095238 68.452381 76.095292 81.571544 0.190476 > 1 - sum(co2 $ resid ^2) / sum((dist - mean(dist))^2)# R^2= 67.4% [1] 0.652418 > > co3 <- cobs(speed, dist, "convex", lambda = -1)# 3 warnings Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. WARNING: Some lambdas had problems in rq.fit.sfnc(): lambda icyc ifl fidel sum|res|_s k [1,] 0.0107117 16 18 263.5 4 18 [2,] 0.3077451 12 18 263.5 4 18 WARNING! Since the optimal lambda chosen by SIC corresponds to the roughest possible fit, you should plot() the returned object (which plots 'sic' against 'lambda') and possibly consider doing one of the following: (1) reduce 'lambda.lo', increase 'lambda.hi', increase 'lambda.length' or all of the above; (2) modify the number of knots. Warning message: In cobs(speed, dist, "convex", lambda = -1) : drqssbc2(): Not all flags are normal (== 1), ifl : 111811111811111111111111111 > summaryCobs(co3) List of 24 $ call : language cobs(x = speed, y = dist, constraint = "convex", lambda = -1) $ tau : num 0.5 $ degree : num 2 $ constraint : chr "convex" $ ic : NULL $ pointwise : NULL $ select.knots : logi TRUE $ select.lambda: logi TRUE $ x : num [1:50] 4 4 7 7 8 9 10 10 10 11 ... $ y : num [1:50] 2 10 4 22 16 10 18 26 34 17 ... $ resid : num [1:50] -5.79 2.21 -9.81 8.19 0 ... $ fitted : num [1:50] 7.79 7.79 13.81 13.81 16 ... $ coef : num [1:20] 7.79 10.6 14.88 17.12 19.55 ... $ knots : num [1:18] 4 7 8 9 10 ... $ k0 : int [1:25] 18 18 18 18 18 18 18 18 15 16 ... $ k : int 3 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 32539 $ lambda : Named num 66.3 ..- attr(*, "names")= chr "lambda" $ icyc : int [1:25] 15 15 16 13 13 12 12 12 14 15 ... $ ifl : int [1:25] 1 1 18 1 1 1 1 18 1 1 ... $ pp.lambda : num [1:25] 0 0.01 0.01 0.02 0.04 0.08 0.16 0.31 0.6 1.18 ... $ pp.sic : num [1:25] 2.37 2.37 2.37 2.37 2.37 ... $ i.mask : logi [1:25] TRUE TRUE TRUE TRUE TRUE TRUE ... cb.lo ci.lo fit ci.up cb.up 1 -18.188896 -9.642853 7.79451 25.2319 33.7779 2 -15.716184 -7.852488 8.19261 24.2377 32.1014 3 -14.435228 -6.860545 8.59486 24.0503 31.6249 4 -14.065626 -6.478846 9.00124 24.4813 32.0681 5 -14.252180 -6.469021 9.41175 25.2925 33.0757 6 -14.667262 -6.611198 9.82641 26.2640 34.3201 7 -15.060722 -6.737504 10.24521 27.2279 35.5511 8 -15.261611 -6.733215 10.66814 28.0695 36.5979 9 -15.161809 -6.525773 11.09521 28.7162 37.3522 10 -14.699858 -6.073933 11.52642 29.1268 37.7527 11 -13.850768 -5.360923 11.96177 29.2845 37.7743 12 -12.622720 -4.392236 12.40126 29.1948 37.4252 13 -11.061185 -3.198385 12.84489 28.8882 36.7510 14 -9.262158 -1.843794 13.29265 28.4291 35.8475 15 -7.397170 -0.443575 13.74455 27.9327 34.8863 16 -5.782277 0.790185 14.20065 27.6111 34.1836 17 -4.506533 1.797751 14.66102 27.5243 33.8286 18 -3.408972 2.687147 15.12568 27.5642 33.6603 19 -2.418466 3.506106 15.59462 27.6831 33.6077 20 -1.580704 4.223984 16.06788 27.9118 33.7165 21 -0.921072 4.824618 16.54814 28.2717 34.0173 22 -0.247350 5.437528 17.03697 28.6364 34.3213 23 0.560677 6.143388 17.53436 28.9253 34.5081 24 1.501406 6.941123 18.04033 29.1395 34.5793 25 2.438439 7.739196 18.55487 29.3706 34.6713 26 3.279139 8.475440 19.07798 29.6805 34.8768 27 4.167235 9.246310 19.60966 29.9730 35.0521 28 5.163885 10.092849 20.14992 30.2070 35.1359 29 6.201031 10.969383 20.69874 30.4281 35.1964 30 7.085760 11.746450 21.25613 30.7658 35.4265 31 7.849515 12.445151 21.82210 31.1990 35.7947 32 8.666999 13.182728 22.39663 31.6105 36.1263 33 9.586054 13.991287 22.97973 31.9682 36.3734 34 10.510397 14.806214 23.57141 32.3366 36.6324 35 11.279572 15.519828 24.17165 32.8235 37.0637 36 12.035693 16.227499 24.78047 33.3334 37.5252 37 12.912159 17.018754 25.39786 33.7770 37.8836 38 13.896917 17.885501 26.02382 34.1621 38.1507 39 14.816824 18.711546 26.65834 34.6051 38.4999 40 15.541399 19.409323 27.30144 35.1936 39.0615 41 16.262211 20.107394 27.95311 35.7988 39.6440 42 17.079027 20.872711 28.61335 36.3540 40.1477 43 17.933019 21.665796 29.28216 36.8985 40.6313 44 18.620298 22.349820 29.95954 37.5693 41.2988 45 19.169812 22.944208 30.64550 38.3468 42.1212 46 19.779172 23.581579 31.34002 39.0985 42.9009 47 20.490014 24.289872 32.04311 39.7963 43.5962 48 21.183490 24.989330 32.75477 40.5202 44.3261 49 21.672603 25.554459 33.47501 41.3956 45.2774 50 22.112296 26.089242 34.20381 42.3184 46.2953 51 22.656042 26.696674 34.94119 43.1857 47.2263 52 23.297520 27.372512 35.68713 44.0018 48.0767 53 23.883744 28.014088 36.44165 44.8692 48.9996 54 24.333295 28.566762 37.20474 45.8427 50.0762 55 24.857974 29.172673 37.97639 46.7801 51.0948 56 25.561042 29.901119 38.75662 47.6121 51.9522 57 26.401263 30.724427 39.54542 48.3664 52.6896 58 27.203642 31.525159 40.34279 49.1604 53.4819 59 27.941406 32.285346 41.14873 50.0121 54.3561 60 28.796171 33.126871 41.96324 50.7996 55.1303 61 29.826683 34.089159 42.78632 51.4835 55.7460 62 30.948462 35.115513 43.61797 52.1204 56.2875 63 31.947912 36.062593 44.45820 52.8538 56.9685 64 32.855747 36.951009 45.30699 53.6630 57.7582 65 33.838848 37.892754 46.16435 54.4360 58.4899 66 34.922331 38.904684 47.03029 55.1559 59.1382 67 35.976908 39.900035 47.90479 55.9095 59.8327 68 36.837596 40.768086 48.78787 56.8076 60.7381 69 37.681155 41.627461 49.67951 57.7316 61.6779 70 38.644794 42.570240 50.57973 58.5892 62.5147 71 39.705763 43.581156 51.48852 59.3959 63.2713 72 40.678958 44.535985 52.40587 60.2758 64.1328 73 41.445084 45.354671 53.33180 61.3089 65.2185 74 42.145882 46.132334 54.26630 62.4003 66.3867 75 42.793009 46.876798 55.20937 63.5419 67.6257 76 43.236599 47.487487 56.16101 64.8345 69.0854 77 43.253419 47.814592 57.12122 66.4278 70.9890 78 43.029423 47.982905 58.09000 68.1971 73.1506 79 42.854469 48.186949 59.06735 69.9478 75.2802 80 42.885261 48.531887 60.05327 71.5747 77.2213 81 43.195018 49.066856 61.04777 73.0287 78.9005 82 43.803006 49.804786 62.05083 74.2969 80.2987 83 44.687239 50.730921 63.06246 75.3940 81.4377 84 45.785830 51.803731 64.08267 76.3616 82.3795 85 46.990994 52.950880 65.11144 77.2720 83.2319 86 48.141987 54.064494 66.14879 78.2331 84.1556 87 49.230317 55.138874 67.19471 79.2505 85.1591 88 50.458163 56.309702 68.24919 80.1887 86.0402 89 51.911812 57.634884 69.31225 80.9896 86.7127 90 53.554788 59.089941 70.38388 81.6778 87.2130 91 55.207203 60.554152 71.46408 82.3740 87.7210 92 56.705814 61.917964 72.55285 83.1877 88.3999 93 58.126881 63.232556 73.65019 84.0678 89.1735 94 59.464570 64.494013 74.75610 85.0182 90.0476 95 60.583736 65.611638 75.87058 86.1295 91.1574 96 61.229633 66.414472 76.99363 87.5728 92.7576 97 61.089584 66.692681 78.12525 89.5578 95.1609 98 59.907545 66.274433 79.26545 92.2565 98.6233 99 57.568406 65.082479 80.41421 95.7459 103.2600 100 54.081906 63.123354 81.57154 100.0197 109.0612 knots : [1] 3.99998 7.00000 8.00000 9.00000 10.00000 11.00000 12.00000 13.00000 [9] 14.00000 15.00000 16.00000 17.00000 18.00000 19.00000 20.00000 22.00000 [17] 23.00000 25.00002 coef : [1] 7.794511 10.595050 14.880952 17.119048 19.547619 22.166667 24.976190 [8] 27.976190 31.166667 34.547619 38.119048 41.880952 45.833333 49.976190 [15] 54.309524 61.095238 68.452381 76.095292 81.571544 0.190476 > 1 - sum(co3 $ resid ^2) / sum((dist - mean(dist))^2) # R^2 = 66.25% [1] 0.652261 > > with(co2, plot(pp.sic ~ pp.lambda, type = "b", col = 3, log = "x")) > with(co3, plot(pp.sic ~ pp.lambda, type = "b", col = 4, log = "x")) > > plot(speed, dist, main = "cobs(speed,dist, ..) for data(cars)") > lines(speed, fitted(co2), col=3); rug(knots(co2), col=3) > lines(speed, fitted(co3), col=4); rug(knots(co3), col=4) > lines(speed, fitted(co1.1), col=2); rug(knots(co1.1), col=2) > detach(cars) > > ##-- another larger example using "random" x > set.seed(101) > x <- round(sort(rnorm(500)), 3) # rounding -> multiple values > sum(duplicated(x)) # 32 [1] 35 > y <- (fx <- exp(-x)) + rt(500,4)/4 > summaryCobs(cxy <- cobs(x,y, "decrease")) qbsks2(): Performing general knot selection ... Deleting unnecessary knots ... List of 24 $ call : language cobs(x = x, y = y, constraint = "decrease") $ tau : num 0.5 $ degree : num 2 $ constraint : chr "decrease" $ ic : chr "AIC" $ pointwise : NULL $ select.knots : logi TRUE $ select.lambda: logi FALSE $ x : num [1:500] -3.18 -2.82 -2.47 -2.32 -2.18 ... $ y : num [1:500] 23.47 17.31 11.75 10.03 8.83 ... $ resid : num [1:500] 3.338 1.739 0.104 -0.198 -0.189 ... $ fitted : num [1:500] 20.13 15.57 11.65 10.23 9.02 ... $ coef : num [1:6] 20.133 4.399 1.931 0.852 0.123 ... $ knots : num [1:5] -3.177 -0.901 -0.329 0.217 2.587 $ k0 : num 6 $ k : num 6 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 2297 $ lambda : num 0 $ icyc : int 20 $ ifl : int 1 $ pp.lambda : NULL $ pp.sic : NULL $ i.mask : NULL cb.lo ci.lo fit ci.up cb.up 1 19.4261539 19.77609785 20.1332535 20.490409 20.840353 2 18.6726860 19.00159764 19.3372876 19.672978 20.001889 3 17.9355059 18.24423770 18.5593319 18.874426 19.183158 4 17.2145986 17.50401040 17.7993864 18.094762 18.384174 5 16.5099479 16.78090751 17.0574512 17.333995 17.604954 6 15.8215360 16.07492016 16.3335261 16.592132 16.845516 7 15.1493446 15.38603897 15.6276112 15.869184 16.105878 8 14.4933543 14.71425424 14.9397066 15.165159 15.386059 9 13.8535457 14.05955619 14.2698122 14.480068 14.686079 10 13.2299004 13.42193541 13.6179279 13.813920 14.005955 11 12.6224016 12.80138349 12.9840539 13.166724 13.345706 12 12.0310364 12.19789391 12.3681901 12.538486 12.705344 13 11.4557979 11.61146321 11.7703365 11.929210 12.084875 14 10.8966884 11.04209248 11.1904931 11.338894 11.484298 15 10.3537226 10.48978919 10.6286599 10.767531 10.903597 16 9.8269318 9.95456916 10.0848369 10.215105 10.342742 17 9.3163680 9.43645864 9.5590241 9.681590 9.801680 18 8.8221075 8.93549614 9.0512215 9.166947 9.280336 19 8.3442530 8.45173361 8.5614292 8.671125 8.778605 20 7.8829348 7.98523676 8.0896470 8.194057 8.296359 21 7.4383082 7.53608415 7.6358751 7.735666 7.833442 22 7.0105499 7.10436496 7.2001134 7.295862 7.389677 23 6.5998510 6.69017570 6.7823618 6.874548 6.964873 24 6.2064091 6.29361624 6.3826205 6.471625 6.558832 25 5.8304201 5.91478545 6.0008894 6.086993 6.171359 26 5.4720696 5.55377713 5.6371685 5.720560 5.802267 27 5.1315260 5.21067635 5.2914578 5.372239 5.451390 28 4.8089343 4.88555627 4.9637573 5.041958 5.118580 29 4.5044102 4.57847544 4.6540670 4.729659 4.803724 30 4.2180355 4.28947509 4.3623870 4.435299 4.506738 31 3.9498516 4.01857620 4.0887171 4.158858 4.227583 32 3.6998525 3.76577568 3.8330574 3.900339 3.966262 33 3.4679740 3.53104114 3.5954080 3.659775 3.722842 34 3.2540798 3.31430374 3.3757688 3.437234 3.497458 35 3.0579437 3.11544919 3.1741397 3.232830 3.290336 36 2.8792298 2.93430782 2.9905209 3.046734 3.101812 37 2.7174793 2.77064796 2.8249123 2.879177 2.932345 38 2.5721207 2.62418085 2.6773139 2.730447 2.782507 39 2.4425238 2.49458826 2.5477257 2.600863 2.652928 40 2.3281087 2.38157725 2.4361477 2.490718 2.544187 41 2.2217481 2.27763180 2.3346671 2.391702 2.447586 42 2.1147630 2.17283231 2.2320983 2.291364 2.349434 43 2.0081703 2.06765255 2.1283606 2.189069 2.248551 44 1.9025658 1.96239353 2.0234542 2.084515 2.144343 45 1.7981929 1.85717812 1.9173789 1.977580 2.036565 46 1.6949971 1.75197884 1.8101348 1.868291 1.925273 47 1.5925922 1.64660058 1.7017220 1.756843 1.810852 48 1.4901187 1.54060923 1.5921403 1.643671 1.694162 49 1.3859782 1.43319747 1.4813898 1.529582 1.576801 50 1.2775940 1.32308124 1.3695059 1.415931 1.461418 51 1.1685486 1.21493954 1.2622865 1.309633 1.356024 52 1.0651401 1.11404012 1.1639478 1.213856 1.262756 53 0.9699195 1.02167151 1.0744900 1.127309 1.179061 54 0.8845413 0.93866941 0.9939130 1.049157 1.103285 55 0.8099473 0.86550951 0.9222168 0.978924 1.034486 56 0.7466077 0.80242934 0.8594013 0.916373 0.972195 57 0.6946637 0.74950019 0.8054667 0.861433 0.916270 58 0.6539655 0.70664640 0.7604129 0.814179 0.866860 59 0.6240094 0.67361353 0.7242399 0.774866 0.824470 60 0.6010731 0.64749825 0.6948801 0.742262 0.788687 61 0.5768335 0.62119253 0.6664657 0.711739 0.756098 62 0.5510070 0.59437589 0.6386386 0.682901 0.726270 63 0.5239391 0.56722289 0.6113987 0.655575 0.698858 64 0.4960558 0.53994871 0.5847461 0.629544 0.673436 65 0.4677904 0.51277214 0.5586809 0.604590 0.649571 66 0.4395281 0.48588777 0.5332028 0.580518 0.626878 67 0.4115797 0.45945260 0.5083121 0.557172 0.605045 68 0.3841775 0.43358397 0.4840087 0.534433 0.583840 69 0.3574825 0.40836321 0.4602925 0.512222 0.563103 70 0.3315950 0.38384096 0.4371636 0.490486 0.542732 71 0.3065642 0.36004204 0.4146220 0.469202 0.522680 72 0.2823954 0.33696919 0.3926677 0.448366 0.502940 73 0.2590553 0.31460556 0.3713006 0.427996 0.483546 74 0.2364749 0.29291629 0.3505209 0.408125 0.464567 75 0.2145513 0.27184943 0.3303284 0.388807 0.446105 76 0.1931496 0.25133684 0.3107232 0.370110 0.428297 77 0.1721053 0.23129540 0.2917053 0.352115 0.411305 78 0.1512286 0.21162924 0.2732746 0.334920 0.395321 79 0.1303116 0.19223336 0.2554313 0.318629 0.380551 80 0.1091385 0.17299880 0.2381752 0.303352 0.367212 81 0.0874987 0.15381915 0.2215064 0.289194 0.355514 82 0.0652000 0.13459734 0.2054248 0.276252 0.345650 83 0.0420805 0.11525159 0.1899306 0.264610 0.337781 84 0.0180160 0.09571916 0.1750236 0.254328 0.332031 85 -0.0070777 0.07595753 0.1607040 0.245450 0.328486 86 -0.0332478 0.05594285 0.1469716 0.238000 0.327191 87 -0.0605107 0.03566683 0.1338264 0.231986 0.328164 88 -0.0888601 0.01513272 0.1212686 0.227404 0.331397 89 -0.1182740 -0.00564846 0.1092981 0.224245 0.336870 90 -0.1487217 -0.02666117 0.0979148 0.222491 0.344551 91 -0.1801682 -0.04788776 0.0871188 0.222125 0.354406 92 -0.2125779 -0.06931016 0.0769101 0.223130 0.366398 93 -0.2459160 -0.09091089 0.0672886 0.225488 0.380493 94 -0.2801505 -0.11267371 0.0582545 0.229183 0.396659 95 -0.3152520 -0.13458383 0.0498076 0.234199 0.414867 96 -0.3511944 -0.15662804 0.0419480 0.240524 0.435090 97 -0.3879545 -0.17879463 0.0346757 0.248146 0.457306 98 -0.4255120 -0.20107332 0.0279907 0.257055 0.481493 99 -0.4638490 -0.22345510 0.0218929 0.267241 0.507635 100 -0.5029500 -0.24593211 0.0163824 0.278697 0.535715 knots : [1] -3.17701 -0.90100 -0.32900 0.21700 2.58701 coef : [1] 20.1332552 4.3994420 1.9310722 0.8518522 0.1225609 0.0163824 > 1 - sum(cxy $ resid ^ 2) / sum((y - mean(y))^2) # R^2 = 95.9% [1] 0.94966 > > ## Interpolation > if(FALSE) { ##-- since it takes too long here! + cpuTime(cxyI <- cobs(x,y, "decrease", knots = unique(x))) + ## takes very long : 1864.46 sec. (Pent. III, 700 MHz) + summaryCobs(cxyI)# only 8 knots remaining! + } > > dx <- diff(range(ux <- unique(x))) > rx <- range(xx <- seq(ux[1] - dx/20, ux[length(ux)] + dx/20, len = 201)) > cpuTime(cxyI <- cobs(x,y, "decrease", knots = ux, nknots = length(ux))) Time elapsed: 0.02 Warning message: In cobs(x, y, "decrease", knots = ux, nknots = length(ux)) : The number of knots can't be equal to the number of unique x for degree = 2. 'cobs' has automatically deleted the middle knot. > ## 17.3 sec. (Pent. III, 700 MHz) > summary(cxyI) COBS regression spline (degree = 2) from call: cobs(x = x, y = y, constraint = "decrease", knots = ux, nknots = length(ux)) {tau=0.5}-quantile; dimensionality of fit: 465 from {465} x$knots[1:464]: -3.17701, -2.82300, -2.46600, ... , 2.58701 coef[1:465]: 23.4711460, 21.8076240, 12.7756454, 11.3273211, 8.8347075, ... , 0.0163826 R^2 = 95.93% ; empirical tau (over all): 245/500 = 0.49 (target tau= 0.5) > pxx <- predict(cxyI, xx) > plot(x,y, cex = 3/4, xlim = rx, ylim = range(y, pxx[,"fit"]), + main = "Artificial (x,y), N=500 : `interpolating' cobs()") > lines(xx, exp(-xx), type = "l", col = "gray40") > lines(pxx, col = "red") > rug(cxyI$knots, col = "blue", lwd = 0.5) > > ## Deg = 1 > cpuTime(cI1 <- cobs(x,y, "decrease", knots= ux, nknots= length(ux), degree = 1)) Time elapsed: 0.02 > summary(cI1) COBS regression spline (degree = 1) from call: cobs(x = x, y = y, constraint = "decrease", knots = ux, nknots = length(ux), degree = 1) {tau=0.5}-quantile; dimensionality of fit: 465 from {465} x$knots[1:465]: -3.17701, -2.82300, -2.46600, ... , 2.58701 coef[1:465]: 23.4711921, 17.3106528, 11.7497490, 10.0279800, 8.8347075, ... , 0.0163826 R^2 = 96.02% ; empirical tau (over all): 251/500 = 0.502 (target tau= 0.5) > pxx <- predict(cI1, xx) > plot(x,y, cex = 3/4, xlim = rx, ylim = range(y, pxx[,"fit"]), + main = paste("Artificial, N=500, `interpolate'", deparse(cI1$call))) > lines(xx, exp(-xx), type = "l", col = "gray40") > lines(pxx, col = "red") > rug(cI1$knots, col = "blue", lwd = 0.5) > > > cpuTime(cxyS <- cobs(x,y, "decrease", lambda = -1)) Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. The algorithm has converged. You might plot() the returned object (which plots 'sic' against 'lambda') to see if you have found the global minimum of the information criterion so that you can determine if you need to adjust any or all of 'lambda.lo', 'lambda.hi' and 'lambda.length' and refit the model. Time elapsed: 0.26 > ## somewhat < 2 sec. (Pent. III, 700 MHz) > pxx <- predict(cxyS, xx) > pxx[xx > max(x) , ]# those outside to the right -- currently all = Inf ! z fit [1,] 2.58988 0.015882677 [2,] 2.62158 0.010364641 [3,] 2.65329 0.004818668 [4,] 2.68499 -0.000755241 [5,] 2.71669 -0.006357087 [6,] 2.74839 -0.011986869 [7,] 2.78009 -0.017644587 [8,] 2.81180 -0.023330242 [9,] 2.84350 -0.029043833 [10,] 2.87520 -0.034785361 > summaryCobs(cxyS) List of 24 $ call : language cobs(x = x, y = y, constraint = "decrease", lambda = -1) $ tau : num 0.5 $ degree : num 2 $ constraint : chr "decrease" $ ic : NULL $ pointwise : NULL $ select.knots : logi TRUE $ select.lambda: logi TRUE $ x : num [1:500] -3.18 -2.82 -2.47 -2.32 -2.18 ... $ y : num [1:500] 23.47 17.31 11.75 10.03 8.83 ... $ resid : num [1:500] 0 0.114 -0.353 -0.338 -0.121 ... $ fitted : num [1:500] 23.47 17.2 12.1 10.37 8.96 ... $ coef : num [1:22] 23.47 8.84 4.31 3.15 2.58 ... $ knots : num [1:20] -3.18 -1.67 -1.29 -1.05 -0.85 ... $ k0 : int [1:25] 21 21 21 21 21 20 20 20 18 18 ... $ k : int 18 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 2297 $ lambda : Named num 0.0201 ..- attr(*, "names")= chr "lambda" $ icyc : int [1:25] 22 26 28 25 26 25 26 25 22 21 ... $ ifl : int [1:25] 1 1 1 1 1 1 1 1 1 1 ... $ pp.lambda : num [1:25] 0 0 0 0.001 0.001 0.003 0.005 0.01 0.02 0.039 ... $ pp.sic : num [1:25] -1.83 -1.83 -1.83 -1.83 -1.83 ... $ i.mask : logi [1:25] FALSE FALSE FALSE FALSE FALSE FALSE ... cb.lo ci.lo fit ci.up cb.up 1 22.60909228 23.19250811 23.4712016 23.749895 24.333311 2 21.54745196 22.09422790 22.3554188 22.616610 23.163386 3 20.51583131 21.02796613 21.2726092 21.517252 22.029387 4 19.51423415 19.99372402 20.2227728 20.451822 20.931311 5 18.54267135 18.99150511 19.2059097 19.420314 19.869148 6 17.60116304 18.02131588 18.2220198 18.422724 18.842877 7 16.68974121 17.08316669 17.2711032 17.459040 17.852465 8 15.80845257 16.17707263 16.3531597 16.529247 16.897867 9 14.95736171 15.30305459 15.4681895 15.633324 15.979017 10 14.13655412 14.46114019 14.6161926 14.771245 15.095831 11 13.34613885 13.65136470 13.7971688 13.942973 14.248199 12 12.58625050 12.87377162 13.0111184 13.148465 13.435986 13 11.85704987 12.12841294 12.2580411 12.387669 12.659032 14 11.15872306 11.41534882 11.5379371 11.660525 11.917151 15 10.49147856 10.73464665 10.8508063 10.966966 11.210134 16 9.85554231 10.08637947 10.1966487 10.306918 10.537755 17 9.25115067 9.47062370 9.5754644 9.680305 9.899778 18 8.67854175 8.88745630 8.9872533 9.087050 9.295965 19 8.13794525 8.33695154 8.4320154 8.527079 8.726086 20 7.62957132 7.81917734 7.9097507 8.000324 8.189930 21 7.15359837 7.33419138 7.4204593 7.506727 7.687320 22 6.71015959 6.88203672 6.9641412 7.046246 7.218123 23 6.29932763 6.46273684 6.5407962 6.618856 6.782265 24 5.92109619 6.07628970 6.1504245 6.224559 6.379753 25 5.57535741 5.72266045 5.7930260 5.863392 6.010695 26 5.26187464 5.40177256 5.4686008 5.535429 5.675327 27 4.97998285 5.11323149 5.1768834 5.240535 5.373784 28 4.71452019 4.84220607 4.9032007 4.964195 5.091881 29 4.45743742 4.58039190 4.6391263 4.697861 4.820815 30 4.20929180 4.32796910 4.3846604 4.441352 4.560029 31 3.97048364 4.08506709 4.1398028 4.194538 4.309122 32 3.74121251 3.85175038 3.9045535 3.957357 4.067895 33 3.52143522 3.62800502 3.6789126 3.729820 3.836390 34 3.31215059 3.41508255 3.4642524 3.513422 3.616354 35 3.11804329 3.21779561 3.2654466 3.313097 3.412850 36 2.93992965 3.03669890 3.0829249 3.129151 3.225920 37 2.77795791 2.87184036 2.9166873 2.961534 3.055417 38 2.63015751 2.72137746 2.7649526 2.808528 2.899748 39 2.48502157 2.57389621 2.6163510 2.658806 2.747680 40 2.34085199 2.42748522 2.4688693 2.510253 2.596887 41 2.19769497 2.28216392 2.3225141 2.362864 2.447333 42 2.06197167 2.14454569 2.1839907 2.223436 2.306010 43 1.93967240 2.02056431 2.0592058 2.097847 2.178739 44 1.82957434 1.90899881 1.9469393 1.984880 2.064304 45 1.72067396 1.79881711 1.8361455 1.873474 1.951617 46 1.61056369 1.68745240 1.7241816 1.760911 1.837799 47 1.50359280 1.57934731 1.6155347 1.651722 1.727477 48 1.41671095 1.49151568 1.5272493 1.562983 1.637788 49 1.34837864 1.42240255 1.4577632 1.493124 1.567148 50 1.26127331 1.33474586 1.3698432 1.404940 1.478413 51 1.14711659 1.22023500 1.2551631 1.290091 1.363210 52 1.05313018 1.12612202 1.1609897 1.195857 1.268849 53 0.99745150 1.07049149 1.1053822 1.140273 1.213313 54 0.93193770 1.00523865 1.0402540 1.075269 1.148570 55 0.82700232 0.90069620 0.9358992 0.971102 1.044796 56 0.71302300 0.78723111 0.8226798 0.858128 0.932337 57 0.63795062 0.71250560 0.7481200 0.783734 0.858289 58 0.60243694 0.67711402 0.7127867 0.748459 0.823136 59 0.59829198 0.67311668 0.7088599 0.744603 0.819428 60 0.59282968 0.66813189 0.7041032 0.740075 0.815377 61 0.58400039 0.66006551 0.6964013 0.732737 0.808802 62 0.57495107 0.65177996 0.6884806 0.725181 0.802010 63 0.56577469 0.64335477 0.6804142 0.717474 0.795054 64 0.54107373 0.61966540 0.6572081 0.694751 0.773342 65 0.47693041 0.55694100 0.5951615 0.633382 0.713393 66 0.40412123 0.48578868 0.5248006 0.563813 0.645480 67 0.34798082 0.43125051 0.4710278 0.510805 0.594075 68 0.30845654 0.39330948 0.4338431 0.474377 0.559230 69 0.28333865 0.36998945 0.4113819 0.452774 0.539425 70 0.26510813 0.35372976 0.3960637 0.438398 0.527019 71 0.25307874 0.34376203 0.3870808 0.430400 0.521083 72 0.24696961 0.33996321 0.3843856 0.428808 0.521802 73 0.23851080 0.33415227 0.3798395 0.425527 0.521168 74 0.22205880 0.32047227 0.3674837 0.414495 0.512909 75 0.19762838 0.29892798 0.3473181 0.395708 0.497008 76 0.16493198 0.26942647 0.3193427 0.369259 0.473753 77 0.12615653 0.23443077 0.2861526 0.337874 0.446149 78 0.09047648 0.20293716 0.2566588 0.310380 0.422841 79 0.05885765 0.17576339 0.2316084 0.287453 0.404359 80 0.03122539 0.15288532 0.2110014 0.269117 0.390777 81 0.00727521 0.13420453 0.1948377 0.255471 0.382400 82 -0.01347494 0.11956518 0.1831175 0.246670 0.379710 83 -0.03271934 0.10753366 0.1745315 0.241529 0.381782 84 -0.05312493 0.09515173 0.1659824 0.236813 0.385090 85 -0.07450384 0.08239138 0.1573391 0.232287 0.389182 86 -0.09673967 0.06929025 0.1486015 0.227913 0.393943 87 -0.11986569 0.05583758 0.1397697 0.223702 0.399405 88 -0.14402704 0.04198646 0.1308437 0.219701 0.405715 89 -0.16944869 0.02766415 0.1218235 0.215983 0.413096 90 -0.19640782 0.01278106 0.1127090 0.212637 0.421826 91 -0.22520954 -0.00276145 0.1035004 0.209762 0.432210 92 -0.25616579 -0.01906423 0.0941974 0.207459 0.444561 93 -0.28957765 -0.03622459 0.0848003 0.205825 0.459178 94 -0.32572171 -0.05433196 0.0753089 0.204950 0.476340 95 -0.36484093 -0.07346487 0.0657233 0.204912 0.496288 96 -0.40713991 -0.09368946 0.0560435 0.205776 0.519227 97 -0.45278411 -0.11505923 0.0462695 0.207598 0.545323 98 -0.50190202 -0.13761570 0.0364012 0.210418 0.574704 99 -0.55458927 -0.16138981 0.0264387 0.214267 0.607467 100 -0.61091376 -0.18640349 0.0163820 0.219167 0.643678 knots : [1] -3.17701 -1.67200 -1.29300 -1.05100 -0.85000 -0.69700 -0.53500 -0.40300 [9] -0.28100 -0.15700 -0.02600 0.16500 0.26900 0.44000 0.54600 0.74700 [17] 0.95200 1.21000 1.55300 2.58701 coef : [1] 23.4712039 8.8370254 4.3078426 3.1485055 2.5789729 2.1356146 [7] 1.8387645 1.5534839 1.4158987 1.1540057 1.0654814 0.7096368 [13] 0.7096368 0.6913826 0.6719781 0.4621872 0.3844119 0.3844119 [19] 0.2057951 0.1061021 0.0163819 11.5718036 > R2 <- 1 - sum(cxyS $ resid ^ 2) / sum((y - mean(y))^2) > R2 # R^2 = 96.3%, now 96.83% [1] 0.956557 > > plot(x,y, cex = 3/4, xlim = rx, ylim = range(y, pxx[,"fit"], finite = TRUE), + main = "Artificial (x,y), N=500 : cobs(*, lambda = -1)") > mtext(substitute(R^2 == r2 * "%", list(r2 = round(100*R2,1)))) > lines(xx, exp(-xx), type = "l", col = "gray40") > lines(pxx, col = "red") > rug(cxyS$knots, col = "blue", lwd = 1.5) > > ## Show print-monitoring : > > cxyS <- cobs(x,y, "decrease", lambda = -1, print.mesg = 2)# << improve! (1 line) Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. fieq=TRUE -> Tnobs = 559, n0 = 59, |ptConstr| = 0 The algorithm has converged. You might plot() the returned object (which plots 'sic' against 'lambda') to see if you have found the global minimum of the information criterion so that you can determine if you need to adjust any or all of 'lambda.lo', 'lambda.hi' and 'lambda.length' and refit the model. > cxyS <- cobs(x,y, "none", lambda = -1, print.mesg = 3) Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. loo.design2(): -> Xeq 501 x 22 (nz = 1501 =^= 0.14%) Xieq 38 x 22 (nz = 152 =^= 0.18%) ........................ The algorithm has converged. You might plot() the returned object (which plots 'sic' against 'lambda') to see if you have found the global minimum of the information criterion so that you can determine if you need to adjust any or all of 'lambda.lo', 'lambda.hi' and 'lambda.length' and refit the model. > > ## this does NOT converge (and "trace = 3" does *not* show it -- improve!) > > cxyC <- cobs(x,y, "concave", lambda = -1) Searching for optimal lambda. This may take a while. While you are waiting, here is something you can consider to speed up the process: (a) Use a smaller number of knots; (b) Set lambda==0 to exclude the penalty term; (c) Use a coarser grid by reducing the argument 'lambda.length' from the default value of 25. WARNING! Since the optimal lambda chosen by SIC rests on a flat portion, you might plot() the returned object (which plots 'sic' against 'lambda') to see if you want to reduce 'lambda.lo' and/or increase 'lambda.ho' > summaryCobs(cxyC) List of 24 $ call : language cobs(x = x, y = y, constraint = "concave", lambda = -1) $ tau : num 0.5 $ degree : num 2 $ constraint : chr "concave" $ ic : NULL $ pointwise : NULL $ select.knots : logi TRUE $ select.lambda: logi TRUE $ x : num [1:500] -3.18 -2.82 -2.47 -2.32 -2.18 ... $ y : num [1:500] 23.47 17.31 11.75 10.03 8.83 ... $ resid : num [1:500] 18.33 12.6 7.47 5.92 4.89 ... $ fitted : num [1:500] 5.14 4.71 4.28 4.1 3.94 ... $ coef : num [1:22] 5.14 4.23 3.09 2.72 2.45 ... $ knots : num [1:20] -3.18 -1.67 -1.29 -1.05 -0.85 ... $ k0 : int [1:25] 21 21 21 21 21 21 21 21 21 21 ... $ k : int 21 $ x.ps :Formal class 'matrix.csr' [package "SparseM"] with 4 slots $ SSy : num 2297 $ lambda : Named num 0.295 ..- attr(*, "names")= chr "lambda" $ icyc : int [1:25] 21 20 16 19 18 21 20 20 20 20 ... $ ifl : int [1:25] 1 1 1 1 1 1 1 1 1 1 ... $ pp.lambda : num [1:25] 0 0 0 0.001 0.001 0.003 0.005 0.01 0.02 0.039 ... $ pp.sic : num [1:25] -0.939 -0.939 -0.939 -0.939 -0.939 ... $ i.mask : logi [1:25] FALSE FALSE FALSE FALSE FALSE FALSE ... cb.lo ci.lo fit ci.up cb.up 1 3.2540132 4.56864309 5.13970311 5.7107631 7.0253930 2 3.3078587 4.53590540 5.06935468 5.6028040 6.8308506 3 3.3554272 4.50126677 4.99900624 5.4967457 6.6425853 4 3.3966923 4.46471925 4.92865781 5.3925964 6.4606233 5 3.4316329 4.42625642 4.85830938 5.2903623 6.2849859 6 3.4602358 4.38587430 4.78796095 5.1900476 6.1156860 7 3.4825004 4.34357268 4.71761252 5.0916524 5.9527246 8 3.4984434 4.29935663 4.64726408 4.9951715 5.7960848 9 3.5081059 4.25323862 4.57691565 4.9005927 5.6457254 10 3.5115613 4.20524084 4.50656722 4.8078936 5.5015732 11 3.5089239 4.15539794 4.43621879 4.7170396 5.3635137 12 3.5003588 4.10375990 4.36587036 4.6279808 5.2313819 13 3.4860907 4.05039475 4.29552192 4.5406491 5.1049532 14 3.4664107 3.99539069 4.22517349 4.4549563 4.9839363 15 3.4416805 3.93885722 4.15482506 4.3707929 4.8679697 16 3.4123305 3.88092472 4.08447663 4.2880285 4.7566227 17 3.3788524 3.82174205 4.01412820 4.2065143 4.6494040 18 3.3417841 3.76147213 3.94377976 4.1260874 4.5457754 19 3.3016889 3.70028555 3.87343133 4.0465771 4.4451738 20 3.2591290 3.63835257 3.80308290 3.9678132 4.3470368 21 3.2146354 3.57583397 3.73273447 3.8896350 4.2508336 22 3.1686751 3.51287123 3.66238604 3.8119008 4.1560970 23 3.1216168 3.44957595 3.59203760 3.7344993 4.0624584 24 3.0736927 3.38601847 3.52168917 3.6573599 3.9696856 25 3.0249556 3.32221480 3.45134074 3.5804667 3.8777258 26 2.9752284 3.25811127 3.38099231 3.5038733 3.7867562 27 2.9240322 3.19356286 3.31064388 3.4277249 3.6972556 28 2.8703441 3.12825982 3.24029544 3.3523311 3.6102468 29 2.8145073 3.06230609 3.16994701 3.2775879 3.5253867 30 2.7573354 2.99594800 3.09959858 3.2032492 3.4418618 31 2.6994314 2.92936823 3.02925015 3.1291321 3.3590689 32 2.6411230 2.86266601 2.95890172 3.0551374 3.2766804 33 2.5823991 2.79583791 2.88855328 2.9812687 3.1947075 34 2.5227652 2.72873426 2.81820485 2.9076754 3.1136445 35 2.4619227 2.66126460 2.74785642 2.8344482 3.0337902 36 2.4004285 2.59359758 2.67750799 2.7614184 2.9545875 37 2.3385189 2.52580476 2.60715956 2.6885143 2.8758002 38 2.2759590 2.45781498 2.53681112 2.6158073 2.7976633 39 2.2125547 2.38956952 2.46646269 2.5433559 2.7203706 40 2.1487596 2.32120569 2.39611426 2.4710228 2.6434689 41 2.0846519 2.25274719 2.32576583 2.3987845 2.5668797 42 2.0198295 2.18407226 2.25541740 2.3267625 2.4910052 43 1.9543928 2.11521127 2.18506896 2.2549267 2.4157452 44 1.8883513 2.04616714 2.11472053 2.1832739 2.3410898 45 1.8217822 1.97696326 2.04437210 2.1117809 2.2669620 46 1.7550253 1.90770248 1.97402367 2.0403449 2.1930220 47 1.6878943 1.83832839 1.90367524 1.9690221 2.1194562 48 1.6202328 1.76879368 1.83332680 1.8978599 2.0464208 49 1.5520736 1.69910823 1.76297837 1.8268485 1.9738831 50 1.4832876 1.62923295 1.69262994 1.7560269 1.9019723 51 1.4139431 1.55918857 1.62228151 1.6853744 1.8306199 52 1.3439885 1.48895940 1.55193308 1.6149067 1.7598776 53 1.2735406 1.41858085 1.48158464 1.5445884 1.6896286 54 1.2025397 1.34803482 1.41123621 1.4744376 1.6199327 55 1.1311646 1.27737546 1.34088778 1.4044001 1.5506110 56 1.0594865 1.20662436 1.27053935 1.3344543 1.4815922 57 0.9881607 1.13597996 1.20019092 1.2644019 1.4122211 58 0.9173004 1.06547651 1.12984248 1.1942085 1.3423845 59 0.8463551 0.99494733 1.05949405 1.1240408 1.2726330 60 0.7745884 0.92416938 0.98914562 1.0541219 1.2037028 61 0.7021077 0.85317521 0.91879719 0.9844192 1.1354867 62 0.6295499 0.78215771 0.84844876 0.9147398 1.0673476 63 0.5569282 0.71112084 0.77810032 0.8450798 0.9992724 64 0.4836357 0.63988083 0.70775189 0.7756230 0.9318681 65 0.4093125 0.56832866 0.63740346 0.7064783 0.8654944 66 0.3343711 0.49658930 0.56705503 0.6375208 0.7997390 67 0.2594524 0.42485683 0.49670660 0.5685564 0.7339608 68 0.1844630 0.35310294 0.42635816 0.4996134 0.6682533 69 0.1089016 0.28117581 0.35600973 0.4308437 0.6031178 70 0.0328311 0.20909450 0.28566130 0.3622281 0.5384915 71 -0.0436111 0.13690066 0.21531287 0.2937251 0.4742368 72 -0.1207780 0.06448731 0.14496444 0.2254416 0.4107069 73 -0.1988521 -0.00820075 0.07461600 0.1574328 0.3480841 74 -0.2774135 -0.08103637 0.00426757 0.0895715 0.2859486 75 -0.3564848 -0.15402645 -0.06608086 0.0218647 0.2243231 76 -0.4365073 -0.22730454 -0.13642929 -0.0455540 0.1636487 77 -0.5180714 -0.30104951 -0.20677772 -0.1125059 0.1045159 78 -0.6008094 -0.37514998 -0.27712616 -0.1791023 0.0465571 79 -0.6844618 -0.44952738 -0.34747459 -0.2454218 -0.0104874 80 -0.7691796 -0.52422741 -0.41782302 -0.3114186 -0.0664664 81 -0.8554282 -0.59939103 -0.48817145 -0.3769519 -0.1209147 82 -0.9439111 -0.67523129 -0.55851988 -0.4418085 -0.1731287 83 -1.0351390 -0.75190283 -0.62886832 -0.5058338 -0.2225976 84 -1.1284587 -0.82920785 -0.69921675 -0.5692256 -0.2699748 85 -1.2235560 -0.90705119 -0.76956518 -0.6320792 -0.3155744 86 -1.3204420 -0.98543625 -0.83991361 -0.6943910 -0.3593852 87 -1.4193598 -1.06443660 -0.91026204 -0.7560875 -0.4011642 88 -1.5207178 -1.14417593 -0.98061048 -0.8170450 -0.4405031 89 -1.6250324 -1.22481061 -1.05095891 -0.8771072 -0.4768855 90 -1.7328771 -1.30651438 -1.12130734 -0.9361003 -0.5097376 91 -1.8448394 -1.38946510 -1.19165577 -0.9938464 -0.5384722 92 -1.9614844 -1.47383392 -1.26200420 -1.0501745 -0.5625240 93 -2.0833282 -1.55977716 -1.33235264 -1.1049281 -0.5813770 94 -2.2108206 -1.64743101 -1.40270107 -1.1579711 -0.5945815 95 -2.3443367 -1.73690906 -1.47304950 -1.2091899 -0.6017623 96 -2.4841766 -1.82830219 -1.54339793 -1.2584937 -0.6026193 97 -2.6305704 -1.92168011 -1.61374636 -1.3058126 -0.5969223 98 -2.7836872 -2.01709401 -1.68409480 -1.3510956 -0.5845024 99 -2.9436449 -2.11457961 -1.75444323 -1.3943068 -0.5652415 100 -3.1105208 -2.21416030 -1.82479166 -1.4354230 -0.5390625 knots : [1] -3.17701 -1.67200 -1.29300 -1.05100 -0.85000 -0.69700 -0.53500 -0.40300 [9] -0.28100 -0.15700 -0.02600 0.16500 0.26900 0.44000 0.54600 0.74700 [17] 0.95200 1.21000 1.55300 2.58701 coef : [1] 5.13970e+00 4.23047e+00 3.09228e+00 2.71711e+00 2.44948e+00 [6] 2.23561e+00 2.04531e+00 1.86770e+00 1.71424e+00 1.56563e+00 [11] 1.41157e+00 1.21704e+00 1.03882e+00 8.72683e-01 7.05337e-01 [16] 5.19867e-01 2.74588e-01 -5.12686e-03 -3.68213e-01 -1.20011e+00 [21] -1.82479e+00 3.65448e-11 > > > dev.off() null device 1 > > proc.time() user system elapsed 3.85 0.25 4.09