{"id":2147,"date":"2020-02-26T00:00:26","date_gmt":"2020-02-25T15:00:26","guid":{"rendered":"http:\/\/141.164.34.82\/?p=2147"},"modified":"2020-02-26T00:13:04","modified_gmt":"2020-02-25T15:13:04","slug":"r-%ed%94%8c%eb%a1%af-%ed%8c%81-2","status":"publish","type":"post","link":"http:\/\/ds.sumeun.org\/?p=2147","title":{"rendered":"R \ud50c\ub86f \ud301 2"},"content":{"rendered":"

\ube45\ub370\uc774\ud130 \ud50c\ub86f\uacfc \uc54c\ud30c<\/h2>\n

\ub370\uc774\ud130\uc758 \ud06c\uae30\uac00 \ub9e4\uc6b0 \ud074 \ub54c, \uae30\ubcf8 \uc0b0\uc810\ub3c4\ub294 \uc5ec\ub7ec \uc810\uc5d0 \uacb9\uce58\uae30 \ub54c\ubb38\uc5d0 \uc54c\ud30c\ub97c \uc870\uc815\ud558\ub294 \ubc29\ubc95\uc744 \ub9ce\uc774 \uc4f4\ub2e4.<\/p>\n

library(data.table)\r\nlibrary(ggplot2)\r\nlibrary(dplyr)\r\n\r\n#N <- 1000\r\nN <- 1000000\r\n\r\nx <- rnorm(N)\r\ny <- x + rnorm(N)\r\n\r\n\r\ndat <- data.table(x = x, \r\n                  y = y)\r\n<\/code><\/pre>\n
dat %>% ggplot(aes(x=x, y=y)) + geom_point()\r\n<\/code><\/pre>\n
\"plot<\/p>\n
dat %>% ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01)\r\n<\/code><\/pre>\n
\"plot<\/p>\n
ggplot2\uc758 \uc54c\ud30c \ucd5c\uc18c\uac12\uc740 0.01<\/code>\uc778 \ub4ef \ud558\ub2e4. \ub530\ub77c\uc11c \ub370\uc774\ud130\uc758 \ud06c\uae30\uac00 \ub9e4\uc6b0 \ud074 \ub54c\uc5d0\ub294 alpha=0.01<\/code>\ub3c4 \ucda9\ubd84\ud788 \uc791\uc9c0 \uc54a\uc740 \uacbd\uc6b0\uac00 \uc788\ub2e4. (\uc704\uc758 \ud50c\ub86f\uc744 \ubcf4\uba74 \uc911\uc559\uc758 \uac80\uc740 \uc0c9\uc740 \ud558\ub098\uc758 \ud070 \ub369\uc5b4\ub9ac\ub85c \ubcf4\uc778\ub2e4. \uc774\uac83\uc740 \ubc00\ub3c4\uac00 \uac70\uc758 \ube44\uc2b7\ud55c \ubd80\ubd84\uc77c \uc218\ub3c4 \uc788\uace0, \ub18d\ub3c4\uac00 \ub354 \uc9d9\uc5b4\uc9c8 \uc218 \uc5c6\uae30 \ub54c\ubb38\uc77c \uc218\ub3c4 \uc788\ub2e4. alpha=1<\/code>\uc5d0\uc11c \uc11c\ub85c \ub18d\ub3c4\uac00 \ubd84\uac04\ub418\uc9c0 \uc54a\ub294 \ubd80\ubd84\uc774 \ud070 \uac83\uacfc \uac19\uc740 \uc774\uc720\uc774\ub2e4.)<\/p>\n
\uc774\ubcc0\ub7c9 \uc815\uaddc\ubd84\ud3ec\ub294 \ub108\ubb34 \ub2e8\uc21c\ud558\ubbc0\ub85c \ucda9\ubd84\ud788 \ubcf5\uc7a1\ud55c \ub370\uc774\ud130\ub97c \uc0ac\uc6a9\ud574\ubcf4\uc790.<\/p>\n
#N <- 100000\r\n#x <- rnorm(N)\r\n#y <- 2*exp(x\/10) + 2*abs(x)*sin(x) + rt(N, df = 30)\r\n\r\nx1 <- rnorm(N\/2)\r\ny1 <- 2*sin(x1) + rnorm(N\/2)\r\nx2 <- rnorm(N\/2)\r\ny2 <- 2*cos(x2) + rt(N\/2, df=30)\r\n\r\ndat <- data.table(x=c(x1,x2),\r\n                  y=c(y1,y2))\r\n<\/code><\/pre>\n
\uc5ec\ub7ec \uc54c\ud30c\uac12<\/h3>\n
\uc774\uc5d0 \ub300\ud55c \ub300\uc548\uc740 \uc5ec\ub7ec \uc54c\ud30c\ub97c \uc0ac\uc6a9\ud558\uc5ec \uc0b0\uc810\ub3c4\ub97c \uadf8\ub9ac\ub294 \uac83\uc774\ub2e4.<\/p>\n
p1 <- dat %>% ggplot(aes(x=x, y=y)) + geom_point() + \r\n  labs(title='alpha=1') + theme_minimal()\r\np2 <- dat %>% ggplot(aes(x=x, y=y)) + geom_point(alpha=0.1) + \r\n  labs(title='alpha=0.1') + theme_minimal()\r\np3 <- dat %>% ggplot(aes(x=x, y=y)) + geom_point(alpha=0.05) +\r\n  labs(title='alpha=0.05') + theme_minimal()\r\np4 <- dat %>% ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) +\r\n  labs(title='alpha=0.01') + theme_minimal()\r\nlibrary(cowplot)\r\n<\/code><\/pre>\n
## \r\n## ********************************************************\r\n<\/pre>\n
## Note: As of version 1.0.0, cowplot does not change the\r\n<\/pre>\n
##   default ggplot2 theme anymore. To recover the previous\r\n<\/pre>\n
##   behavior, execute:\r\n##   theme_set(theme_cowplot())\r\n<\/pre>\n
## ********************************************************\r\n<\/pre>\n
plot_grid(p1,p2,p3,p4,ncol=2)\r\n<\/code><\/pre>\n
\"plot<\/p>\n
\ud558\uc9c0\ub9cc \uc54c\ud30c\uc758 \ucd5c\uc800\uac12 alpha=0.01<\/code>\ub97c \uc368\ub3c4 \uc911\uc2ec\ubd80\uc758 \ubc00\ub3c4 \ucc28\uc774\uac00 \uc790\uc138\ud788 \ub4dc\ub7ec\ub098\uc9c0 \uc54a\ub294\ub2e4. \uc774\ub7f0 \uacbd\uc6b0 \uc0d8\ud50c\ub9c1\uc744 \ud65c\uc6a9\ud560 \uc218 \uc788\ub2e4.<\/p>\n
\uc0d8\ud50c\ub9c1<\/h3>\n
p1 <- dat %>% sample_n(N\/5) %>% \r\n  ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) + \r\n  labs(title='alpha=0.01 with 20% of data') + theme_minimal()\r\np2 <- dat %>% sample_n(N\/10) %>% \r\n  ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) + \r\n  labs(title='alpha=0.01 with 10% of data') + theme_minimal()\r\np3 <- dat %>% sample_n(N\/50) %>% \r\n  ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) +\r\n  labs(title='alpha=0.01 with 2% of data') + theme_minimal()\r\np4 <- dat %>% sample_n(N\/100) %>% \r\n  ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) +\r\n  labs(title='alpha=0.01 with 1% of data') + theme_minimal()\r\nlibrary(cowplot)\r\nplot_grid(p1,p2,p3,p4,ncol=2)\r\n<\/code><\/pre>\n
\"plot<\/p>\n
\ud558\uc9c0\ub9cc \uc0d8\ud50c\ub9c1\uc740 \uc804\uccb4 \ub370\uc774\ud130\uc758 \uc77c\ubd80\ub9cc\uc744 \uc0ac\uc6a9\ud558\uace0 \uc788\uace0, \ubb34\uc791\uc704\ub85c \uc0d8\ud50c\uc744 \ubf51\uace0 \uc788\uae30 \ub54c\ubb38\uc5d0 \ub370\uc774\ud130\uc758 \ubc00\ub3c4\ub97c \uc815\ud655\ud558\uac8c \ub098\ud0c0\ub0b4\uc9c0 \ubabb\ud55c\ub2e4\ub294 \ub2e8\uc810\uc774 \uc788\ub2e4.<\/p>\n
\uc870\uac74\ubd80 \ubd84\ud3ec<\/h2>\n
\uc0b0\uc810\ub3c4\ub97c \uadf8\ub824\ubcf4\ub294 \uc774\uc720\ub294 \uc5ec\ub7ec \uac00\uc9c0\uac00 \uc788\ub2e4. \ud55c \uac00\uc9c0 \uc774\uc720\ub294 \uc120\ud615\ud68c\uadc0\ubd84\uc11d, ML, DL \ub4f1\uc744 \ud1b5\ud574 \ud68c\uadc0 \ubd84\uc11d\uc744 \ud558\uae30 \uc804\uc5d0 \uc801\ud569\ud55c \ubaa8\ud615\uc744 \ud655\uc778\ud558\uae30 \uc704\ud574 EDA\ub97c \ud558\ub294 \uac83\uc774\ub2e4.<\/p>\n
\uc774\ub54c \uc911\uc694\ud55c \uac83\uc740 \uc774\ubcc0\ub7c9 \ubd84\ud3ec\ubcf4\ub2e4\ub294 \uc870\uac74\ubd80 \ubd84\ud3ec, \uc989 \\(\\mathbb{p}(y|x)\\) \ub97c \ud655\uc778\ud558\ub294 \uac83\uc774\ub2e4. \uc774\uc5d0 \ubc18\ud574, \uc704\uc758 \ubaa8\ub4e0 \ud50c\ub86f\uc740 \uc774\ubcc0\ub7c9 \ubd84\ud3ec\uc5d0 \ucd1b\uc810\uc744 \ub9de\ucd94\uace0 \uc788\ub2e4.<\/p>\n
\ub9cc\uc57d \\(x\\) \uc870\uac74\ubd80 \ub300\ud45c\uac12\uc744 \ud655\uc778\ud558\uace0\uc790 \ud55c\ub2e4\uba74 \ub2e4\uc74c\uacfc \ube44\ubaa8\uc218 \ud68c\uadc0\uc120\uc744 \uadf8\ub824\ubcfc \uc218 \uc788\ub2e4.<\/p>\n
\ud68c\uadc0\uc120<\/h3>\n
p1 <- dat %>% sample_n(N\/100) %>% \r\n  ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) +\r\n  geom_smooth(method='loess') + \r\n  labs(title='alpha=0.01 with 1% of data, loess') + theme_minimal() \r\np2 <- dat %>% sample_n(N\/100) %>% \r\n  ggplot(aes(x=x, y=y)) + geom_point(alpha=0.01) +\r\n  geom_smooth(method='auto') + \r\n  labs(title='alpha=0.01 with 1% of data, gam') + theme_minimal() \r\nprint(p1)\r\nprint(p2)\r\n<\/code><\/pre>\n
## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = \"cs\")'\r\n<\/pre>\n
\"plot\"plot<\/p>\n
\uc774 \ubc29\ubc95\uc744 \ud1b5\ud574 \uc870\uac74\ubd80 \ud3c9\uade0( \\(\\mathbb{E}[y|x] = \\int y\\ \\mathbb{p}(y|x) dy\\) )\uc744 \ud655\uc778\ud560 \uc218 \uc788\uc9c0\ub9cc, \uc870\uac74\ubd80 \ubd84\ud3ec\uae4c\uc9c0 \ud655\uc778\ud558\uae30\ub294 \uc5b4\ub835\ub2e4.<\/p>\n
\uc870\uac74\ubd80 \ubd84\ud3ec<\/h3>\n
\\(x\\) \uad6c\ud68d<\/h4>\n
\uc55e\uc758 \uc608\uc5d0\uc11c \ubd24\ub4ef\uc774 \uc77c\uad04\uc801\uc73c\ub85c \uc54c\ud30c\ub97c \uc870\uc815\ud558\ub294 \ubc29\ubc95\uc740 \ub370\uc774\ud130\uac00 \ubc00\uc9d1\ub418\uc5b4 \uc788\ub294 \uacf3\uc5d0\uc11c\ub294 \uc5ec\uc804\ud788 \ub18d\ub3c4\uac00 \ucd5c\ub300\uac12\uc774 \ub418\uc5b4 \ubc00\ub3c4\ub97c \uad6c\ubd84\ud560 \uc218 \uc5c6\uc73c\uba70, \ub370\uc774\ud130\uac00 \ub4ec\uc131\ud55c \uacf3\uc5d0\uc11c\ub294 \uc810\uc774 \ub108\ubb34 \ud750\ub824\uc11c \ub370\uc774\ud130\ub97c \ud655\uc778\ud560 \uc218 \uc5c6\ub2e4. \ud55c \uac00\uc9c0 \ubc29\ubc95\uc740 \\(x\\) \ub97c \uad6c\uac04\ubcc4\ub85c \uad6c\ud68d\ud574\uc11c \uc801\uc808\ud788 \uc0d8\ud50c\ub9c1\ud558\ub294 \uac83\uc774\ub2e4. \ub2e4\uc74c\uacfc \uac19\ub2e4.<\/p>\n
dat %>% \r\n  mutate(xCut = cut(x, breaks=10)) %>%\r\n  group_by(xCut) %>%\r\n  do(sample_n(., 10000, replace=TRUE)) %>%\r\n  ggplot(aes(x=x, y=y)) + \r\n  geom_point(alpha=0.01)\r\n<\/code><\/pre>\n
\"plot<\/p>\n
\ud558\uc9c0\ub9cc \ub370\uc774\ud130\uc758 \ubc00\ub3c4\uac00 \\(x\\) \ubc29\ud5a5\uc73c\ub85c \ubb3c\uacb0\uce58\ub294 \ub4ef\uc774 \ubcf4\uc778\ub2e4. \uc774\ub294 \uad6c\ud68d \uad6c\uac04\uc774 \ub113\uc5b4\uc11c \uc0dd\uae30\ub294 \ubd80\uc791\uc6a9\uc774\ub2e4. \uad6c\ud68d \uad6c\uac04\uc744 \uc904\uc778\ub2e4\uba74 \uc774\ub7f0 \ubb34\ub2ac\ub294 \uc798 \ubcf4\uc774\uc9c0 \uc54a\ub294\ub2e4.<\/p>\n
dat %>% \r\n  mutate(xCut = cut(x, breaks=50)) %>%\r\n  group_by(xCut) %>%\r\n  do(sample_n(., 2000, replace=TRUE)) %>%\r\n  ggplot(aes(x=x, y=y)) + \r\n  geom_point(alpha=0.01)\r\n<\/code><\/pre>\n
\"plot<\/p>\n
\\(x\\) \ubc00\ub3c4 \ucd94\uc815<\/h4>\n
\\(x\\) \uc5d0 \uc778\uc758\uc801\uc778 \uad6c\ud68d\uc744 \ud558\uae30 \uc2eb\ub2e4\uba74, \\(x\\) \uc758 \ubc00\ub3c4\ub97c \ucd94\uc815\ud560 \uc218\ub3c4 \uc788\uc744 \uac83\uc774\ub2e4. \uadf8\ub9ac\uace0<\/p>\n
xDensity <- ks::kde(dat$x)\r\ndat$prob <- predict(xDensity, x = dat$x)\r\n\r\n#head(dat)\r\ndat %>% \r\n  mutate(xCut = cut(x, breaks=50)) %>%\r\n  group_by(xCut) %>%\r\n  do(sample_n(., 2000, replace=TRUE, weight=1\/prob)) %>%\r\n\r\n  ggplot(aes(x=x, y=y)) + \r\n  geom_point(alpha=0.01) \r\n<\/code><\/pre>\n
\"plot<\/p>\n
\\(x\\) \ubd84\ud3ec \uac10\uc548<\/h2>\n
\ub9c8\uc9c0\ub9c9\uc73c\ub85c \ubaa8\ub4e0 \\(x\\) \ub97c \ub3d9\uc77c\ud558\uac8c \ucde8\uae09\ud558\ub294 \uac83\uc774 \ubb38\uc81c\ub77c\uace0 \uc0dd\uac01\ud55c\ub2e4\uba74, \\(x\\) \uc758 \ubc00\ub3c4\ub97c \uc57d\uac04 \uac10\uc548\ud558\uc5ec(\ubaa8\ub450 \uac10\uc548\ud55c\ub2e4\uba74 \uc0b0\uc810\ub3c4\uc640 \ub3d9\uc77c\ud558\uac8c \ub418\ubbc0\ub85c, \ubc00\ub3c4\uc5d0 \\(\\log\\) \ub97c \ucde8\ud558\ub358\uc9c0 \ud574\uc11c) \uc0b0\uc810\ub3c4\ub97c \uadf8\ub824\ubcfc \uc218 \uc788\uc744 \uac83\uc774\ub2e4.<\/p>\n","protected":false},"excerpt":{"rendered":"
\ube45\ub370\uc774\ud130 \ud50c\ub86f\uacfc \uc54c\ud30c \ub370\uc774\ud130\uc758 \ud06c\uae30\uac00 \ub9e4\uc6b0 \ud074 \ub54c, \uae30\ubcf8 \uc0b0\uc810\ub3c4\ub294 \uc5ec\ub7ec \uc810\uc5d0 \uacb9\uce58\uae30 \ub54c\ubb38\uc5d0 \uc54c\ud30c\ub97c \uc870\uc815\ud558\ub294 \ubc29\ubc95\uc744 \ub9ce\uc774 \uc4f4\ub2e4. library(data.table) library(ggplot2) library(dplyr) #N <- 1000 N <- 1000000 x <- rnorm(N) y <- x + rnorm(N) dat <- data.table(x = x, y = y) dat %>% ggplot(aes(x=x, y=y)) + geom_point() dat %>% ggplot(aes(x=x, y=y)) + […]<\/p>\n","protected":false},"author":1,"featured_media":2149,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[165],"tags":[],"jetpack_featured_media_url":"http:\/\/ds.sumeun.org\/wp-content\/uploads\/2020\/02\/RplotTips02.png","_links":{"self":[{"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts\/2147"}],"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=2147"}],"version-history":[{"count":2,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts\/2147\/revisions"}],"predecessor-version":[{"id":2150,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/posts\/2147\/revisions\/2150"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=\/wp\/v2\/media\/2149"}],"wp:attachment":[{"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2147"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2147"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/ds.sumeun.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2147"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}