{"id":1954,"date":"2019-09-01T22:55:16","date_gmt":"2019-09-01T13:55:16","guid":{"rendered":"http:\/\/141.164.34.82\/?p=1954"},"modified":"2019-09-02T00:10:39","modified_gmt":"2019-09-01T15:10:39","slug":"black-box-%eb%aa%a8%ed%98%95%ec%9d%84-%ec%82%b4%ed%8e%b4%eb%b3%b4%ea%b8%b0-%eb%aa%a8%ed%98%95%ec%9d%98-%ec%98%88%ec%b8%a1-%ea%b3%bc%ec%a0%81%ed%95%a9-%ec%9e%94%ec%b0%a8-01","status":"publish","type":"post","link":"http:\/\/ds.sumeun.org\/?p=1954","title":{"rendered":"Black Box \ubaa8\ud615\uc744 \uc0b4\ud3b4\ubcf4\uae30: \ubaa8\ud615\uc758 \uc608\uce21, \uacfc\uc801\ud569, \uc794\ucc28 01"},"content":{"rendered":"

\uc608\uce21 \ubaa8\ud615 \uc0b4\ud3b4\ubcf4\uae30<\/h1>\n
library(dplyr)\nlibrary(ggplot2)\nlibrary(randomForest)\n<\/code><\/pre>\n
\uc608\uce21 \ubaa8\ud615\uc740 \uae30\ubcf8\uc801\uc73c\ub85c \\(f(\\vec{x})\\) \uc774\ub2e4.[1]<\/sup> \uc8fc\uc5b4\uc9c4 \uc124\uba85 \ubcc0\uc218 \\(\\vec{x}\\) \uc5d0 \ub300\ud574 \uacb0\uacfc \ubcc0\uc218 \\(y\\) \ub97c \uc608\uce21\ud55c\ub2e4. \uc608\uce21 \ubaa8\ud615\uc744 \ud3c9\uac00\ud558\ub294 \ubc29\ubc95\uc740 \\(\\hat{y}=f(\\vec{x})\\) \ub77c\uace0 \ud560 \ub54c, \\(|y – \\hat{y}|\\) \uc744 \uc4f8 \uc218\ub3c4 \uc788\uace0, \\((y – \\hat{y})^2\\) \ub97c \uc4f0\uae30\ub3c4 \ud558\uace0, hinge loss\ub77c\ub294 \uac83\uc744 \uc4f0\uae30\ub3c4 \ud558\uc9c0\ub9cc, \uc5b4\uca0b\ub4e0 \uc8fc\uc5b4\uc9c4 \ub370\uc774\ud130\ub97c \uac00\uc7a5 \uc798 \uc124\uba85<\/strong>\ud558\uba74\uc11c\ub3c4 \uc77c\ubc18\ud654<\/strong>\ub97c \uc798 \ud558\ub294 \\(f(\\vec{x})\\) \ub97c \ub9cc\ub4e4\uc5b4\ub0b4\ub294 \uac83\uc774\ub2e4. <\/p>\n
[1]:<\/sup> \ubaa8\uc9d1\ub2e8\uc5d0\uc11c \\((y – \\hat{y})^2\\) \uc744 \ucd5c\uc801\ud654\ud558\ub294 \ud568\uc218 \\(f(x)\\) \uc640 \uc774 \ud568\uc218\uc758 \ucd94\uc815\ud568\uc218 \\(\\hat{f(x)}\\) \ub97c \uad6c\ubd84\ud558\uae30 \uc704\ud574 \uc608\uce21 \ubaa8\ud615\uc744 \\(\\hat{f(x)}\\) \ub85c \ub098\ud0c0\ub0b4\uae30\ub3c4 \ud55c\ub2e4. <\/p>\n
\uc608\uc81c\ub97c \uc0b4\ud3b4\ubcf4\uae30\uc5d0 \uc55e\uc11c TRUE<\/code>\ub97c \ubcf4\uc790. \ub098\ub294 \uac00\ub054 \ud0c0\uc774\ud551\uc744 \ube60\ub974\uac8c \ud558\ub2e4\ubcf4\uba74 TRUE<\/code>\ub97c TURE<\/code>\ub85c \uc4f0\ub294 \uc2e4\uc218\ub97c \uc655\uc655[2]<\/sup> \ud55c\ub2e4. \ud558\ub3c4 \uc790\uc8fc \ud558\ub2e4 \ubcf4\ub2c8 \uadf8\ub0e5 \ub2e4\uc74c\uacfc \uac19\uc774 \uc124\uc815\ud558\ub294\uac8c \ub0ab\uaca0\ub2e4 \uc2f6\ub2e4.<\/p>\n
[2]:<\/sup> \uc655\uc655: \ubd80\uc0ac  \uc2dc\uac04\uc758 \uac04\uaca9\uc744 \ub450\uace0 \uc774\ub530\uae08.<\/p>\n
TURE <- TRUE\n<\/code><\/pre>\n
\uc608\uc81c<\/h2>\n
dat<\/code>\uc5d0\ub294 \ubaa8\ub450 10\uac1c\uc758 \ubcc0\uc218\uac00 \uc788\ub2e4. <\/p>\n
\uc2e4\uc81c \ub370\uc774\ud130 \uc0dd\uc131 \ubaa8\ud615\uc740 \ub2e4\uc74c\uacfc \uac19\ub2e4. \uc774\ub54c \\(x_3\\) \uc740 \\(-1\\), \\(+1\\) \uac12\ub9cc \uac00\uc9c8 \uc218 \uc788\uace0, \ub098\uba38\uc9c0 \ubcc0\uc218\ub4e4\uc740 \ubaa8\ub450 \uc5f0\uc18d\ud615\uc774\ub2e4.<\/p>\n
\\[y = x_1^2 + \\frac{x_2}{|x_1|+0.1} + \\Big(I(x_2<0)+I(x_2<1)\\times x_3\\Big) \\times \\frac{1}{5}(x_4+x_5+x_6+x_7+x_8)+e\\]<\/p>\n
\\[e\\sim\\mathcal{N}(0,1)\\]<\/p>\n
genData = function(N=1000, iseed=100) {\n  set.seed(iseed)\n  sdErr = 1\n\n  x1 <- runif(N, -3, 3)\n  x11 <- x1^2 \n\n  x2 <- runif(N, -3, 3)\n  x3 <- sample(c(-1, 1), N, replace=TRUE)\n\n  ncolMatX <- 5\n  ncolMatX2 <- 2\n\n  require(mvtnorm)\n  sigma <- matrix(\n    rep(0.97, ncolMatX*ncolMatX),\n    ncolMatX, ncolMatX)\n  diag(sigma) = 1\n\n  matX <- rmvnorm(N, mean=rep(0,ncolMatX), sigma=sigma)\n  x40 <- apply(matX, 1, mean)\/ncolMatX\n  colnames(matX) <- paste0('x', 4:(ncolMatX+3))\n\n  matX2 <- rmvnorm(N, mean=rep(0,ncolMatX2), sigma=diag(ncolMatX2))\n  colnames(matX2) <- paste0('x', (ncolMatX+4):(ncolMatX+ncolMatX2+3))\n\n  y <- x11 + x2\/(abs(x1)+0.1) + (x2 < 0)*x40 + (x2<1)*x3 * x40 + rnorm(N, 0, sdErr)\n\n  return(data.frame(y, x1, x2, x3, matX, matX2))\n}\n\ndat <- genData()\n<\/code><\/pre>\n
dat<\/code>\uc5d0\ub294 \uacb0\uacfc \ubcc0\uc218 y<\/code>\uc640 \uc124\uba85\ubcc0\uc218 (\ud6c4\ubcf4) x1<\/code> ~ x10<\/code>\uc774 \uc800\uc7a5\ub418\uc5b4 \uc788\ub2e4. \uacb0\uacfc \ubcc0\uc218 y<\/code>\uc640 \uc124\uba85\ubcc0\uc218 x1<\/code> ~ x10<\/code>\uc758 \uad00\uacc4\ub294 \ub2e4\uc74c\uacfc \uac19\ub2e4.<\/p>\n
pairs(dat[,c("y", "x1", "x2", "x3")])\npairs(dat[,c("y", paste0("x", 4:10))])\n<\/code><\/pre>\n
\"plot\"plot<\/p>\n
\uc774 \ub370\uc774\ud130\uc5d0 \ub300\ud55c \uc608\uce21 \ubaa8\ud615\uc740 \ub2e4\uc591\ud55c \ubc29\ubc95\uc73c\ub85c \ub9cc\ub4e4 \uc218 \uc788\ub2e4. <\/p>\n
fitLM <- lm(y ~ ., data=dat) ## linear model\n\nlibrary(randomForest)\nfitRF <- randomForest(y ~ ., data = dat, importance = TRUE)\n\nlibrary(gbm) # install.packages('gbm')\nfitGBM <- gbm(y ~ ., data=dat, dist="gaussian", inter=2, n.trees=1000)\n\nlibrary(gam) ## gam # install.packages('gam')\nfitGAM <- gam(y ~ s(x1)+s(x2)+x3+s(x4)+s(x5)+s(x6)+s(x7)+s(x8)+s(x9)+s(x10),\n              data=dat) # x3 is a categorical variable\n\nlibrary(nnet) ## nnet\nset.seed(4)\nfitNN10 <- nnet(y ~ ., data=dat,\n                linout=TRUE, size=10, decay=0.01, maxit=1000, \n                trace=FALSE)\n#fitNN20 <- nnet(y ~ ., linout=TRUE, size=20, data=dat)\nfitNN40 <- nnet(y ~ ., linout=TRUE, size=40, data=dat, maxit=1000, trace=FALSE)\n#fitNN80 <- nnet(y ~ ., linout=TRUE, size=80, data=dat)\nfitNN160 <- nnet(y ~ ., data=dat, \n                 linout=TRUE, size=160, \n                 MaxNWts= 10000, trace=FALSE,\n                 decay=0.01, maxit=10000)\n<\/code><\/pre>\n
(\ud788\ub4e0 160\uac1c \uc9dc\ub9ac 3-\uce35 \uc778\uacf5\uc2e0\uacbd\ub9dd\uc740 \uc9c4\uc9dc \ud558\uc138\uc6d4[3]<\/sup> \uc774\ub2e4!)<\/p>\n
[3]:<\/sup> \uc5b8\uc81c\uc77c\uc9c0 \uac00\ub2a0\ud558\uae30 \ud798\ub4e0 \uc815\ub3c4\uc758 \uc138\uc6d4\uc774\ub098 \uae30\ud55c.<\/p>\n
\uc77c\ub2e8 \ub79c\ub364\ud3ec\uc2a4\ud2b8 \ubaa8\ud615\uc5d0\uc11c \ubcc0\uc218 \uc911\uc694\ub3c4\ub97c \ubf51\uc544\ubcf4\uc790. <\/p>\n
print(fitRF)\n<\/code><\/pre>\n
## \n## Call:\n##  randomForest(formula = y ~ ., data = dat, importance = TRUE) \n##                Type of random forest: regression\n##                      Number of trees: 500\n## No. of variables tried at each split: 3\n## \n##           Mean of squared residuals: 5.280945\n##                     % Var explained: 70.19\n<\/pre>\n
varImpPlot(fitRF)\n<\/code><\/pre>\n
\"plot<\/p>\n
cond<\/strong>itional vis<\/strong>ualization for statistical models : condvis<\/code><\/h2>\n
\ud328\ud0a4\uc9c0 condvis<\/code>\ub294 \uc870\uac74\ubd80 \uc2dc\uac01\ud654(cond<\/strong>itional vis<\/strong>ualziation)\uc758 \uc57d\uc790\uc774\ub2e4. \ub2e4\ub978 \ubcc0\uc218\ub4e4\uc744 \uc77c\uc815\ud55c \uac12(\uc5f0\uc18d\ud615\uc5d0\uc11c \uc911\uc559\uac12, \ubc94\uc8fc\ud615\uc5d0\uc11c \ucd5c\ube48\uac12)\uc73c\ub85c \uace0\uc815\ud55c \ud6c4 \uac00\uc7a5 \uc601\ud5a5\ub825\uc774 \ud070 \ubcc0\uc218(x1<\/code>)\uc758 \ubcc0\ud654\uc5d0 \ub530\ub978 \uacb0\uacfc\ubcc0\uc218\uc758 \ubcc0\ud654\ub97c \ubcf4\uace0\uc790 \ud55c\ub2e4\uba74, \ub2e4\uc74c\uacfc \uac19\uc774 \uc4f8 \uc218 \uc788\ub2e4.<\/p>\n
model<\/code>\uc758 \uacb0\uacfc\ubcc0\uc218(response<\/code>) \ucd08\uace1\uba74\uc5d0\uc11c \ud2b9\uc815\ud55c \ubcc0\uc218(sectionvars<\/code>)\uc758 \ubcc0\ud654\ub97c \ubcf4\uc5ec\uc904 \uc218 \uc788\ub294 \ub2e8\uba74\uc744 \uadf8\ub824\uc900\ub2e4. \ub9cc\uc57d \ubc30\uacbd\ubcc0\uc218\uc758 \uac12\uc744 \ubcc0\uacbd\ud558\uace0 \uc2f6\ub2e4\uba74 \uc624\ub978\ucabd\uc758 \ubd84\ud3ec\uc5d0\uc11c \ud074\ub9ad\ud558\uc5ec \uc870\uac74\uac12\uc744 \ubcc0\uacbd\ud560 \uc218 \uc788\ub2e4. <\/p>\n
# ceplot1d_01, ceplot1d_02\nlibrary(condvis)\nceplot(data=dat, \n       model=fitRF, \n       response="y", \n       sectionvars = "x1")\n\nceplot(data=dat, \n       model=fitGBM, \n       response="y", \n       sectionvars = "x1")\n<\/code><\/pre>\n
\"\"\/
\n\"\"\/<\/p>\n
\ub450 \ubcc0\uc218\uc758 \ubcc0\ud654\ub97c \ud0d0\uc0c9\ud574 \ubcfc \uc218\ub3c4 \uc788\ub2e4. <\/p>\n
#ceplot2d_01, ceplot2d_02\nrequire(condvis)\nceplot(data=dat, \n       model=fitRF, \n       response="y", \n       sectionvars = c("x1", "x2"))\n\nceplot(data=dat, \n       model=fitGBM, \n       response="y", \n       sectionvars = c("x1", "x2"))\n<\/code><\/pre>\n
\"\"\/<\/p>\n
#ceplot2d_03, ceplot2d_04\nceplot(data=dat, \n       model=fitNN10, \n       response="y", \n       sectionvars = c("x1", "x2"), \n       threshold = 1,\n       view3d=TRUE)\n<\/code><\/pre>\n
\"\"\/<\/p>\n
\uc774 \ubc16\uc5d0\ub3c4 \ubaa8\ud615\uc744 \ud0d0\uc0c9\ud558\ub294 \ub2e4\ub978 \ubc29\ubc95\ub4e4\uc774 \uc874\uc7ac\ud558\uba70, \ud328\ud0a4\uc9c0\ub85c\ub3c4 \uad6c\uc131\ub418\uc5c8\ub2e4. \ud2b9\ud788 \uc5b4\ub5bb\uac8c \uc608\uce21\uc774 \ub418\ub294\uc9c0 \ubc16\uc5d0\uc11c \uc27d\uac8c \uc54c \uc218 \uc5c6\ub294 Black Box \ubaa8\ud615\uc744 \uc774\ud574\ud558\ub294\ub370 \ub3c4\uc6c0\uc774 \ub41c\ub2e4. <\/p>\n
p<\/strong>artial-d<\/strong>ependence p<\/strong>lot(pdp<\/code>)<\/h2>\n
p<\/strong>artial-d<\/strong>ependence p<\/strong>lot\ub294 \uc8fc\uc694 \uad00\uc2ec\uc774 \ub418\ub294 \ubcc0\uc218\uc640 \ub098\uba38\uc9c0 \ubcc0\uc218(\ubc30\uacbd\ubcc0\uc218)\ub85c \ubcc0\uc218\ub97c \uad6c\ubd84\ud55c \ud6c4, \ubc30\uacbd \ubcc0\uc218\uc5d0 \ub300\ud574\uc11c\ub294 \ud3c9\uade0<\/strong>\uc744 \ucde8\ud558\ub294 \ubc29\uc2dd\uc774\ub2e4. <\/p>\n
\uc774\ub54c \uc911\uc694\ud55c \uac83\uc740 \ubc30\uacbd \ubcc0\uc218\ub97c \ud3c9\uade0\ud558\ub294 \uac83\uc774 \uc544\ub2c8\ub77c, \ub098\uba38\uc9c0 \ubcc0\uc218\ub4e4\uc5d0 \ub300\ud574 \uacb0\uacfc \ubcc0\uc218\ub97c \ud3c9\uade0\ud558\ub294 \uac83\uc774\ub2e4. <\/p>\n
\uc608\ub97c \ub4e4\uc5b4 \\(x_1\\) \uc5d0 \uad00\uc2ec\uc774 \uc788\uace0, \\(x_2\\) \uac00 \ubc30\uacbd\ubcc0\uc218\ub77c\uba74, \ub370\uc774\ud130\uc5d0\uc11c \\(x_1\\) \uc774 \ubcc0\ud568\uc5d0 \ub530\ub77c \\(f(\\vec{x})\\) \uc758 \ud3c9\uade0\uc774 \uc5b4\ub5bb\uac8c \ubcc0\ud558\ub294\uc9c0 \ud655\uc778\ud558\ub294 \uac83\uc774\ub2e4. \uc218\uc2dd\uc73c\ub85c \uc4f0\uba74 \ub2e4\uc74c\uacfc \uac19\ub2e4.(\uc544\ub798\uc5d0\uc11c \\(\\vec{x}\\) \ub294 \uc804\uccb4 \uc124\uba85 \ubcc0\uc218 \uc911 \\(x_1\\) \ub97c \uc81c\uc678\ud55c \uc124\uba85\ubcc0\uc218\uc774\ub2e4.)<\/p>\n
\\[\\mathbb{E}_{\\vec{x} \\sim p(\\vec{x}|x_1)} f(x_1, \\vec{x})\\]<\/p>\n
\ubb3c\ub860 \uc6b0\ub9ac\ub294 \\(p(\\vec{x}|x_1)\\) \ub3c4 \ubaa8\ub974\uace0, \\(f\\) \ub3c4 \ubaa8\ub974\ubbc0\ub85c \ub370\uc774\ud130\ub97c \ud3c9\uade0\ud574\uc11c \uadf8\ub9b0\ub2e4. <\/p>\n
randomForest::partialPlot(fitRF, pred.data=dat, x.var='x1')\nrandomForest::partialPlot(fitRF, pred.data=dat, x.var='x2')\n<\/code><\/pre>\n
\"plot\"plot<\/p>\n
pdp<\/code> \ud328\ud0a4\uc9c0\ub97c \ud65c\uc6a9\ud558\uba74 \ub2e4\ub978 \ubaa8\ud615\uc5d0 \ub300\ud574\uc11c\ub3c4 pdp(p<\/strong>artial-d<\/strong>ependence p<\/strong>lot)\ub97c \uadf8\ub9b4 \uc218 \uc788\ub2e4.<\/p>\n
\ubcc0\uc218 \ud558\ub098\ub97c \ubcc0\ud654\uc2dc\ud0ac \uc218\ub3c4 \uc788\uace0, \ub450 \uac1c\ub97c \ubcc0\ud654\uc2dc\ud0ac \uc218\ub3c4 \uc788\ub2e4. <\/p>\n
library(pdp)\nplotPartial(partial(fitRF, pred.var='x1', train=dat),\n            main='Random Forest')\nplotPartial(partial(fitGAM, pred.var='x1', train=dat),\n            main='GAM')\nplotPartial(partial(fitGBM, pred.var='x1', train=dat,\n                    n.trees=1000), main='GBM')\nplotPartial(partial(fitNN10, pred.var='x1', train=dat),\n            main='NN with 10 hidden nodes')\nplotPartial(partial(fitNN160, pred.var='x1', train=dat),\n            main='NN with 160 hidden nodes')\n<\/code><\/pre>\n
\"plot\"plot\"plot\"plot\"plot<\/p>\n
#library(doParallel)\n#cl <- makeCluster(3)\n#registerDoParallel(cl)\n<\/code><\/pre>\n
require(pdp)\nplotPartial(partial(fitRF, pred.var=c('x1','x2'), \n                    train=dat, levelplot=TRUE), \n            main='Random Forest') # , parallel=TRUE\nplotPartial(partial(fitGAM, pred.var=c('x1','x2'), \n                    train=dat, levelplot=TRUE), \n            main='GAM') \nplotPartial(partial(fitGBM, pred.var=c('x1','x2'), \n                    train=dat, n.trees=1000, levelplot=TRUE),\n            main='GBM')\nplotPartial(partial(fitNN10, pred.var=c('x1','x2'), \n                    train=dat, levelplot=TRUE), \n            main='NN w\/ 10 hidden nodes')\nplotPartial(partial(fitNN160, pred.var=c('x1','x2'), \n                    train=dat, levelplot=TRUE), \n            main='NN w\/ 160 hidden nodes')\n<\/code><\/pre>\n
\"plot\"plot\"plot\"plot\"plot<\/p>\n
\ud788\ud2b8 \ub9f5\uc774 \uc544\ub2c8\ub77c 3\ucc28\uc6d0 \uadf8\ub9bc\uc744 \ubcf4\uace0 \uc2f6\ub2e4\uba74 \ub2e4\uc74c\uacfc \uac19\uc774 \uc4f4\ub2e4. <\/p>\n
plotPartial(partial(fitLM, \n                    pred.var=c('x1','x2'), \n                    train=dat), \n            levelplot=FALSE, zlab = "y",\n            main = 'Linear Regression Model') \n\nplotPartial(partial(fitRF, \n                    pred.var=c('x1','x2'), \n                    train=dat), \n            levelplot=FALSE, zlab = "y",\n            main = 'Random Forest') \n\nplotPartial(partial(fitGAM, \n                    pred.var=c('x1','x2'), \n                    train=dat, n.trees=1000), \n            levelplot=FALSE, zlab = "y",\n            main = 'GAM') \n\nplotPartial(partial(fitGBM, \n                    pred.var=c('x1','x2'), \n                    train=dat, n.trees=1000), \n            levelplot=FALSE, zlab = "y",\n            main = 'GBM') \n\nplotPartial(partial(fitNN10, \n                    pred.var=c('x1','x2'), \n                    train=dat), \n            levelplot=FALSE, zlab = "y",\n            main = 'NN w\/ 10 hidden nodes')\n\nplotPartial(partial(fitNN160, \n                    pred.var=c('x1','x2'), \n                    train=dat), \n            levelplot=FALSE, zlab = "y",\n            main = 'NN w\/ 160 hidden nodes') \n<\/code><\/pre>\n
\"plot\"plot\"plot\"plot\"plot\"plot<\/p>\n
i<\/strong>ndividual c<\/strong>onditional e<\/strong>xpectation : ice = TRUE<\/code><\/h2>\n
pdp<\/code> \ud328\ud0a4\uc9c0\uc758 partial<\/code>\ub97c \ud65c\uc6a9\ud558\uc5ec ice<\/strong>(i<\/strong>ndividual c<\/strong>onditional e<\/strong>xpectation)\uc744 \ucd9c\ub825\ud560 \uc218\ub3c4 \uc788\ub2e4. \uc774\ub294 \uc804\uccb4\ub97c \ub370\uc774\ud130\uc5d0 \ub300\ud574 \ud3c9\uade0\ud55c \uac12\uc774 \uc544\ub2c8\ub77c, \uac01\uac01 \ub370\uc774\ud130 \ud3ec\uc778\ud2b8\uc5d0 \ub300\ud574 \uad00\uc2ec\uc788\ub294 \ubcc0\uc218\ub97c \ubcc0\ud654\uc2dc\ucf30\uc744 \ub54c \uacb0\uacfc \ubcc0\uc218\uac00 \uc5b4\ub5bb\uac8c \uc608\uce21\ub418\ub294\uac00\ub97c \uadf8\ub824\ubcf8 \uac83\uc774\ub2e4. <\/p>\n
plotPartial(partial(fitRF, \n                    pred.var='x1', \n                    ice=TURE, train=dat), \n            alpha=0.1, main='Random Forest')\nplotPartial(partial(fitGAM, \n                    pred.var='x1', \n                    ice=TRUE, train=dat), \n            alpha=0.1, main='GAM')\n#plotPartial(partial(fitGBM, \n#                    pred.var='x1', \n#                    ice=TRUE, train=dat, n.trees=1000),\n#            alpha=0.1, main='GBM')\nplotPartial(partial(fitNN40, \n                    pred.var='x1', ice=TRUE, train=dat),\n            alpha=0.1, main='NN w\/ 40 hidden nodes')\n<\/code><\/pre>\n
\"plot\"plot\"plot<\/p>\n
#stopCluster(cl)\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"
\uc608\uce21 \ubaa8\ud615 \uc0b4\ud3b4\ubcf4\uae30 library(dplyr) library(ggplot2) library(randomForest) \uc608\uce21 \ubaa8\ud615\uc740 \uae30\ubcf8\uc801\uc73c\ub85c \\(f(\\vec{x})\\) \uc774\ub2e4.[1] \uc8fc\uc5b4\uc9c4 \uc124\uba85 \ubcc0\uc218 \\(\\vec{x}\\) \uc5d0 \ub300\ud574 \uacb0\uacfc \ubcc0\uc218 \\(y\\) \ub97c \uc608\uce21\ud55c\ub2e4. \uc608\uce21 \ubaa8\ud615\uc744 \ud3c9\uac00\ud558\ub294 \ubc29\ubc95\uc740 \\(\\hat{y}=f(\\vec{x})\\) \ub77c\uace0 \ud560 \ub54c, \\(|y – \\hat{y}|\\) \uc744 \uc4f8 \uc218\ub3c4 \uc788\uace0, \\((y – \\hat{y})^2\\) \ub97c \uc4f0\uae30\ub3c4 \ud558\uace0, hinge loss\ub77c\ub294 \uac83\uc744 \uc4f0\uae30\ub3c4 \ud558\uc9c0\ub9cc, \uc5b4\uca0b\ub4e0 \uc8fc\uc5b4\uc9c4 \ub370\uc774\ud130\ub97c \uac00\uc7a5 \uc798 \uc124\uba85\ud558\uba74\uc11c\ub3c4 \uc77c\ubc18\ud654\ub97c […]<\/p>\n","protected":false},"author":1,"featured_media":1957,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[371,146,319],"tags":[373,368,375,338,369,374],"jetpack_featured_media_url":"http:\/\/ds.sumeun.org\/wp-content\/uploads\/2019\/09\/vs_condvis.png","_links":{"self":[{"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts\/1954"}],"collection":[{"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1954"}],"version-history":[{"count":3,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts\/1954\/revisions"}],"predecessor-version":[{"id":1959,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts\/1954\/revisions\/1959"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/media\/1957"}],"wp:attachment":[{"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1954"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1954"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1954"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}