Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Rcpp for everyone
Search
teuder
March 31, 2017
Technology
1
1k
Rcpp for everyone
Brief introduction of learning resources for Rcpp.
teuder
March 31, 2017
Tweet
Share
More Decks by teuder
See All by teuder
あまぁいRcpp生活
teuder
1
460
Other Decks in Technology
See All in Technology
『星の世界の地図の話: Google Sky MapをAI Agentでよみがえらせる』 - Google Developers DevFest Tokyo 2025
taniiicom
0
400
【ASW21-02】STAMP/CAST分析における生成AIの支援 ~羽田空港航空機衝突事故を題材として (Support of Generative AI in STAMP/CAST Analysis - A Case Study Based on the Haneda Airport Aircraft Accident -)
hianraku9498
0
210
adk-samples に学ぶデータ分析 LLM エージェント開発
na0
3
750
不確実性に備える ABEMA の信頼性設計とオブザーバビリティ基盤
nagapad
4
8.4k
ローカルVLM OCRモデル + Gemini 3.0 Proで日本語性能を試す
gotalab555
1
190
AS59105におけるFreeBSD EtherIPの運用と課題
x86taka
0
280
ABEJA FIRST GUIDE for Software Engineers
abeja
0
3.2k
確実に伝えるHealth通知 〜半自動システムでほどよく漏れなく / JAWS-UG 神戸 #9 神戸へようこそ!LT会
genda
0
150
経営から紐解くデータマネジメント
pacocat
4
1.1k
AI駆動開発2025年振り返りとTips集
knr109
1
120
個人から巡るAI疲れと組織としてできること - AI疲れをふっとばせ。エンジニアのAI疲れ治療法 ショートセッション -
kikuchikakeru
5
1.9k
Excelデータ分析で学ぶディメンショナルモデリング ~アジャイルデータモデリングへ向けて~ by @Kazaneya_PR / 20251126
kazaneya
PRO
3
320
Featured
See All Featured
How to train your dragon (web standard)
notwaldorf
97
6.4k
jQuery: Nuts, Bolts and Bling
dougneiner
65
8k
Learning to Love Humans: Emotional Interface Design
aarron
274
41k
The World Runs on Bad Software
bkeepers
PRO
72
12k
The Illustrated Children's Guide to Kubernetes
chrisshort
51
51k
The Invisible Side of Design
smashingmag
302
51k
Fight the Zombie Pattern Library - RWD Summit 2016
marcelosomers
234
17k
Scaling GitHub
holman
464
140k
Building a Modern Day E-commerce SEO Strategy
aleyda
45
8.1k
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
36
6.1k
A better future with KSS
kneath
239
18k
How STYLIGHT went responsive
nonsquared
100
5.9k
Transcript
3DQQGPSFWFSZPOF ବ@teuder
"CPVUNF ాਅथ .BTBLJ&5TVEB 1I% -JGFTDJFODFT #VTJOFTT$POTVMUBOU XJUI %BUB4DJFODF
3DQQ 3DQQJTBQBDLBHFUIBUFOBCMFTZPV UPJNQMFNFOU3GVODUJPOTJO$ #ZVTJOH3DQQ ZPVDBOXSJUF$ DPEFJOBTUZMFTJNJMBSUP3
None
%JSL&EEFMCVFUUFM
5PEBZ *EPOPUUBMLBCPVUIPXUPVTF3DQQ CVUJOUSPEVDFZPVTPNFVTFGVMSFTPVSDFT XIFOZPVXBOUUPMFBSO3DQQ
5IFNBKPSSFTPVSDFTGPS MFBSOJOH3DQQ "EWBODFE3 3DQQ(BMMFSZ 3DQQ/PUF 7JHOFUUFT
"EWBODFE3 (SFBUJOUSPEVDUJPOUP3DQQ #VUJUMBDLTTPNFEFUBJMT )PXDBO*NPWFUPUIFOFYUTUFQ
3DQQ(BMMFSZ (SFBUDPNQJMBUJPOPGHPPEFYBNQMFT #VUUIFJOGPSNBUJPOJTGSBHNFOUFE )PXDBO*HFUUJEZLOPXMFEHFBCPVU3DQQ
3DQQ/PUF (SFBU$ TUZMFSFGFSFODFEPDVNFOU #VUJUPOMZQSPWJEFMJNJUFEJOGPSNBUJPO )PXDBO*HFUNPSFQSBDUJDBM BOENPSFTZTUFNBUJDBMLOPXMFEHF
*OUIFTFEBZT *DPVMEOPUpOEBEPDVNFOUMJLFUIBU*XBOU #VUPOFEBZ *SFNFNCFSFEBQISBTF *`WFTFFOPOUIFJOUFSOFU
If there does not exist what you want, just create
it. That should have been the oath of HOXO-M ! ͳ͚Ε࡞Δɻ ͦΕ͕ϗΫιΤϜͷ͍ͩͬͨͣʂ
IUUQTXXXHJUCPPLDPNCPPLUFVEFSJOUSPEVDUJPOUPSDQQEFUBJMT *XSPUFB3DQQ8FC#PPL
&OHMJTIWFSTJPOJTBMTPBWBJMBCMF IUUQTXXXHJUCPPLDPNCPPLUFVEFSJOUSPEVDUJPOUPSDQQEFUBJMT
$POUFOUTPGUIFCPPL 7FDUPS .BUSJY -JTU %BUB'SBNF 4 3MJLFGVODUJPOT 1SPCBCJMJUZEJTUSJCVUJPOT 6TJOH3GVODUJPOTGSPN3DQQ "UUSJCVUFT
/"/BO*OG &OWJSPONFOU &SSPSIBOEMJOH BOENPSFʜ
4JODFUIF&OHMJTIWFSTJPO XBTSFMFBTFE 1BHFWJFXTEPVCMFEBOEJTTUJMMHSPXJOH 3FMFBTFPG&OHWFS /PW 5JNFTFSJFTPGQBHFWJFXTJOUIFMBTUNPOUIT
1BHFWJFXTCZDPVOUSJFT (FSNBOZJTUIFSEQMBDF #VUMBSHFMZCFIJOEGSPN64BOE+BQBO 1BHFWJFXTJOUIFMBTUNPOUIT
8IBUXPVMECF UIFHPPETJUVBUJPOUPVTF3DQQ w -PPQTJOXIJDIMBUFSJUFSBUJPOEFQFOETPOUIF QSFWJPVTJUFSBUJPO w 3FDVSSFOUDBMMTPGGVODUJPOJOMPPQ w "DDFTTJOHFBDIFMFNFOUTPGWFDUPSEBUBGSBNF w
$IBOHJOHTJ[FPGBWFDUPSBUSVOUJNF w 0QFSBUJPOUIBUOFFEBEWBODFEEBUBTUSVDUVSF BOEBMHPSJUIN
#FDBVTF3DQQSFRVJSFTZPVBOFYUSBDPTU :PVNJHIUCFBCMFUPNBYJNJ[FZPVSRVBMJUZPGMJGF XIFOZPVVTFJUUPCSFBLCPUUMFOFDLPGZPVSDPEF 8IBUXPVMECF UIFHPPETJUVBUJPOUPVTF3DQQ
*ONZDBTF $POWFSUJOHEBUBGSBNFUPTQBSTFNBUSJY df %>% as.matrix %>% Matrix::Matrix(sparse = TRUE) #VUDPOWFSUJOHNBUSJYGBJMTXIFOUIFEBUBJT
SFMBUJWFMZMBSHF BCPVUTFWFSBMUFOTPG(# df %>% asSparseMatrix 4P*DSFBUFB3DQQGVODUJPODPOWFSUJOHUIFEBUB EJSFDUMZUPTQBSTFNBUSJY
#include <Rcpp.h> using namespace Rcpp; %FpOJOHBGVODUJPO // [[Rcpp::export]] S4 asSparseMatrix(
DataFrame df ){ // prerequisite : // all the elements DataFrame is numeric/integer // and not containing NAs. // number of rows and columns int nrow = df.nrows(); int ncol = df.length();
6TJOHTUEWFDUPSJOTUFBEPG3DQQ7FDUPS #FDBVTFDIBOHJOHTJ[FPGWFDUPSBUSVOUJNF JTOPUFGpDJFOUJO3DQQ7FDUPS std::vector<R_xlen_t> rows; std::vector<R_xlen_t> cols; std::vector<double> vals;
4UPSJOHUIFQPTJUJPOTBOEWBMVFTPG OPO[FSPFMFNFOUTJOUIF%BUB'SBNF for(R_xlen_t col=0; col<ncol; ++col){ NumericVector column = df[col]; for(R_xlen_t row=0; row<nrow; ++row){ if(column[row]!=0.0){ rows.push_back(row+1); cols.push_back(col+1); vals.push_back(column[row]); } } }
$BMMJOH.BUSJYTQBSTF.BUSJY Environment env = Environment::namespace_env("Matrix"); Function sparseMatrix = env["sparseMatrix"];
$POWFSUJOHTUEWFDUPSUP/VNFSJD7FDUPS S4 sm = sparseMatrix( Named("i") = wrap(rows), Named("j") = wrap(cols), Named("x") = wrap(vals), Named("dims") = NumericVector::create(nrow,ncol)); 4FUUJOHSPXOBNFTBOEDPMOBNFT List dimnames = List::create(R_NilValue, df.names()); sm.attr("Dimnames") = dimnames; 3FUVSOJOHUIFTQBSTFNBUSJY return sm; }
$PODMVTJPO w :PVDBOJNQMFNFOU3GVODUJPOJO$ XJUI3DQQ w 5IFSFBSFTV⒏DJFOUSFTPVSDFTGPSMFBSOJOH3DQQ w 3DQQJTTVJUBCMFUPCSFBLCPUUMFOFDLTPGZPVSDPEF *IPQFNZl*OUSPEVDUJPOUP3DQQz XPVMEBJECSPBESBOHFPGQFPQMF
&OKPZ
"EWFSUJTJOH 605 1 2 3 4 5 6 7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 24 ষɹɹRcpp Rcpp RcppRͷؔΛC++Ͱ࣮Ͱ͖ΔύοέʔδͰ͢ɻRͱྨࣅͨ͠ελΠϧͰهड़Ͱ͖ ΔΑ͏ʹ࣮͞Ε͍ͯΔͨΊɺC++ʹਂ͍͕ࣝͳͯ͘ར༻͘͢͠ͳ͍ͬͯ·͢ɻ͠ ͔ɺͦͷͨΊͷ࣮ߦ٘ਜ਼ʹ͞Ε͍ͯͳ͍ͷͰɺ୭ͰϋΠύϑΥʔϚϯεͳ݁ՌΛ ಘΔ͜ͱ͕Ͱ͖·͢ɻ 24-1 Rcppͷ׆༻γʔϯ ࣍ͷΑ͏ͳέʔεC++Ͱ࣮͢Δ͜ͱʹΑΓɺRͱൺͯߴԽ͕ݟࠐΊ·͢ɻ ɾ ܁Γฦ͠ॲཧɺಛʹ࣍ͷॲཧ͕લͷॲཧʹґଘ͓ͯ͠ΓฒྻԽͰ͖ͳ͍ ɾ ϕΫ τϧߦྻͷݸʑͷཁૉΞΫηε͢Δඞཁ͕͋Δ ɾ ϕΫ τϧͷαΠζΛಈతʹมߋ͍ͨ͠ ɾ ߴͳσʔλߏΞϧΰϦζϜΛ༻͍ͨॲཧΛߦ͍͍ͨ RcppͷύϑΥʔϚϯεΛࣔͨ͢Ίɺ܁Γฦ͠ॲཧͷྫͱͯ͠MCMCΞϧΰϦζϜͷҰछͰ͋ ΔΪϒεαϯϓϥʔ ʢ1ʣ ͷ࣮ྫΛࣔ͠·͢ɻ͜ͷྫͰɺΪϒεαϯϓϥʔʹΑΓඪ४2ม ਖ਼ن͔ΒnαϯϓϦϯά͍ͯ͠·͢ɻ ·ͣൺֱͷͨΊRΛ༻͍࣮ͨྫΛࣔ͠·͢ ʢϦετ24.1ʣ ɻ Ϧε τ24.1ɹGibbs.R GibbsR <- function(b, n, t){ # 2มඪ४ਖ਼ن͔ΒnαϯϓϦ ϯά # b : 2มͷڞࢄ # n : αϯϓϧ # t : αϯϓϦ ϯάࣺͤͣͯΔִؒ X <- matrix(0, nrow = n, ncol = 2) x1 <- x2 <- 0 sd <- sqrt(1-b^2) for(i in 1:n){ for(j in 1:t){ x1 <- rnorm(1, b*x2, sd) x2 <- rnorm(1, b*x1, sd) ʢ1ʣ ߴ࣍ݩͷෳࡶͳ֬ʹै͏ཚΛੜ͢ΔϚϧίϑ࿈ϞϯςΧϧϩ๏ͱݺΕΔΞϧΰϦζϜͷҰछͰ͢ɻ 24-1 24 ষ ٕज़ධࣾɹ੫ࠐԁ