tag:blogger.com,1999:blog-24082283Sat, 06 Dec 2025 07:31:10 +0000RTeXMiscBUGSTravelBayesian MethodsConferenceGraphicsResearch MethodStataStatisticsICPSRLinuxCCausal Inferencehttp://yusung.blogspot.com/noreply@blogger.com (Yu-Sung Su)Blogger124125tag:blogger.com,1999:blog-24082283.post-1386930836838145834Wed, 22 Sep 2010 15:55:00 +00002010-09-22T11:55:59.647-04:00Misc前天在開會時，系上提到了要開一門電影與政治，又提到了一部最近國內很夯的電影「山楂樹之戀」，結果我做了我自己都不相信的事，去電影院看國片，更甚者，同時間上映的有Inception這部好片，我是入境隨俗吧，一句話短評，淡淡的哀傷！對，就只是淡淡的！ 以下有雷，怕被雷到的，請按Ctrl + W關閉視窗吧！ 先說觀眾吧，果不其然，年輕人都去看Inception了，我周圍坐下的都是頭髮雪白的老公公老婆婆們，也許是特地來看他們那個時代的故事吧。 這是一部以文革為時代背景的電影，對於那個年代陌生的我來說，很多場景在我看來都是好笑的，除此之外，這部電影就圍繞在男女主角的互動上，而這一切都那麼平淡，平淡的讓全場觀眾好些時候都彷彿停住呼吸一般靜默！ 我最有意見的莫過於它每個章節的過門，都用上字幕來提醒觀眾發生了什麼事，這樣粗糙的過門手法，本不應該發生在這個年代的電影上，但細想，這一整部片不就是傳達著平淡、http://yusung.blogspot.com/2010/09/blog-post.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-6824660119682966843Thu, 01 Jul 2010 16:47:00 +00002010-07-01T12:51:05.690-04:00Misc這幾天都在想個問題.....「要是我幫忙了他，是不是這件事就不會發生？」(Had I helped him, would this not happen?) &nbsp;但是這個問題的另一面向，更令人玩味：「是不是當初不幫忙他，這件事就會發生？」(Would this happen had I not helped him?)　這整個問題的關鍵其實不在於幫不幫忙，而是這無可預知已經發生的事What would have happened? 學了好幾年的因果推論，我深深了解到因果推論的無力感，既便任何事的發生有其因果，我們仍然無法確定現在的結果是來自我們覺得的原因。事情的發生無法重來，所以我們沒有法子像實驗室做實驗一樣找出真正的因，What would have&nbsp;happened&nbsp;had I done this? 我們總以為幫助人是有益的，所以伸手幫忙，但是有時候結果http://yusung.blogspot.com/2010/07/blog-post.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-6515038528492262252Thu, 01 Jul 2010 15:46:00 +00002010-07-01T11:46:29.081-04:00Travel寫這篇之前，正在看以前吉他社學姊陳綺貞「旅行的意義」的MV，以前就好欣賞這位學姊，可惜入社後，卻從來沒機會見到她，OK，重點是，看完後，突然了解到有時候人追求不到的人、事、物，都藏在潛意識裡，於是見到類似的風景叫似曾相識 Deja vu，碰到對的人叫一見如故，甚至一見鍾情，這次來到法國也是為了尋找曾經的Deja vu吧。 開完UseR會後，我刻意在行程計畫裡，安排自己住到市中心，以就近接下來去Mt Michel 及St Malo搭車之處，找到了一間便宜的旅館，叫Hotel Campanile， 沒有無線上網，但房間剛裝修好小了點但是很乾淨（它又靠近TGV火車站，方便我前往巴黎）。 其實到了旅館後，已是中午，肚子餓得很，卻因為語言問題，一直不知道該吃什麼好，後來才想起來曾經在路上看過旅遊指南上法國麥當勞連鎖店「Quick」，對於只會有限法文的我，點餐夠用了，它又剛好http://yusung.blogspot.com/2010/07/rennes.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-6448051832430872078Thu, 24 Jun 2010 17:42:00 +00002010-06-24T13:42:40.661-04:00TeXI found that control coloring in TeX is not an easy task. &nbsp;Sometimes, you want your font to be colored as brown. &nbsp;But brown is not explicitly defined in the color package. &nbsp;I found a web page that provides detailed color pallet with the color name and the CMYK and RGB setting. Here is the link to the webpage. http://www.december.com/html/spec/colorcodes.html So you can look up http://yusung.blogspot.com/2010/06/cmyk-and-rgb-color-code-in-tex_24.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-8957320116723083952Sun, 06 Jun 2010 19:26:00 +00002010-06-06T15:38:11.861-04:00BUGSLinuxRI have an entry here about how to install JAGS and rjags in Fedora. &nbsp;However, I have exchanged emails with Robert Dorazio and Dr. Martyn Plummer. &nbsp;Here are the easier steps to compile JAGS and install rjags in R. 1. After you extract the source code of JAGS, go to the folder where the JAGS makefile is using a terminal. cd JAGS-2.1.0 2. In a terminal, type the following sudo ./configurehttp://yusung.blogspot.com/2010/06/general-steps-to-compile-jags-in-linux.htmlnoreply@blogger.com (Yu-Sung Su)4tag:blogger.com,1999:blog-24082283.post-8687777021762506991Sun, 06 Jun 2010 19:08:00 +00002010-06-06T15:09:26.146-04:00BUGSLinuxRChristopher has the RPMs (for Fedora users) for JAGS, coda, and rjags on his blog. The specific blog entry is here: http://cddesjardins.wordpress.com/2010/06/04/jags-package-for-fedora-13-64-bit/ If you have trouble installing JAGS or rjags in Fedora, this is going to simplify the installation process.http://yusung.blogspot.com/2010/06/jags-rjags-and-coda-packages-for-fedora.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-5586526591123946067Wed, 13 Jan 2010 16:05:00 +00002010-07-01T11:47:04.378-04:00I have been converted myself from a Word user to a TeX user for over 5 years. &nbsp;Since the release of xeTeX (a new TeX package that can use system fonts to type Non-Egnlish [UTF-8] documents), I have been using it whenever I wanted to type a Chinese document. &nbsp;However, it is not easy. &nbsp;The biggest hurdle is that you have to write some commands in the&nbsp;preamble&nbsp;and have http://yusung.blogspot.com/2010/01/texing-chinese-japanese-korean-utf-8.htmlnoreply@blogger.com (Yu-Sung Su)2tag:blogger.com,1999:blog-24082283.post-5665529182370007914Mon, 30 Nov 2009 21:16:00 +00002009-12-07T12:00:41.870-05:00MiscRHere is an OpenSSH software I personally found useful and working well under Windows 7 / Vista. &nbsp;It is copSSH. Besides it is free and open source, it is still actively maintained. &nbsp;It works well with the UAC (user account control setting) of Windows 7 / Vista. &nbsp;And it does not clash with Rtools.exe. http://yusung.blogspot.com/2009/11/openssh-software-for-windows-7-vista.htmlnoreply@blogger.com (Yu-Sung Su)1tag:blogger.com,1999:blog-24082283.post-9006931426661865887Sat, 14 Nov 2009 08:25:00 +00002009-11-14T03:48:55.292-05:00TeXThe current WinEdt needs some adjustment to work with MikTeX 2.8. The developer of WinEdt has a post here to show how to do this. I tried to follow his instruction. But it still took me some times to figure out which file I need to make the change. Basically, we need to alter the "menu.dat." The file is located in:Windows XPC:\Program Files\WinEdt Team\WinEdt\Config\MikTeX\menu.datWindows http://yusung.blogspot.com/2009/11/miktex-28-and-winedt-56.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-4913674860269713448Tue, 03 Nov 2009 21:25:00 +00002009-11-22T13:01:59.821-05:00Travel大學時的法國行，就唯獨法國西北方我沒去過，所以出發前，我google map看了好幾次，也從台灣帶了旅行指南，但是就因為supper shuttle來得太早，我匆忙下就忘了旅行指南，這也就註定了計畫趕不上變化的開始。 飛機到了法國戴高樂機場，法國的海關檢查真是迅速，一分鐘我就通關了，領了行李，我就按計畫找TGV高鐵前往Rennes。個性小心的我，在這個交通轉換時間，安排了四個多小時的空檔，我可以慢慢的去尋找十多年前還未存在的機場TGV車站，然後慢慢去了解月台在那，把玩一台又一台的車票販費機，記清所有的動線，為回程做準備。 當TGV入站時，我的回憶漸漸回來了，TGV裡餐車貴的誇張的三明治更是令我記憶深刻，我大概清醒不到半小時，就一路睡到了Rennes，下車時真正的緊張來了。 我找尋去會場的地鐵站，當然第一件事就是買票，我記得出發前，書上說法國近年來信用卡全面改成晶片卡，http://yusung.blogspot.com/2009/11/rennes.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-6597392121958578025Mon, 26 Oct 2009 20:40:00 +00002009-10-26T16:56:24.596-04:00MiscThis semester, I moved down to the office in the 9th floor. I shared the office with Matt. We like to open the office door so that the air circulation is better. But our office seems to be the only one on the 9th floor that remains open most of the time. If you wonder why others don't open the door, here is the reason: people walk in and ask you questions unrelated to your expertise. Here http://yusung.blogspot.com/2009/10/our-office-becomes-service-counter.htmlnoreply@blogger.com (Yu-Sung Su)2tag:blogger.com,1999:blog-24082283.post-7944289310268164949Tue, 04 Aug 2009 06:24:00 +00002009-11-03T16:28:03.088-05:00Travel從法國回來都二個多禮拜了，我再不寫寫心得可能就寫不出來了。先來個一小段，希望這次不要無疾而終！會動念想去法國，除了剛好有個useR的學術會議在那舉行，我可以堂而皇之的、師出有名的去之外，就是因為十幾年前跟陳學長自助旅行那段回憶。為了那次的法國行，我學了半年的法文，我們足跡幾乎踏遍法國及周圍國家，但是總覺得回憶就像我的法文一樣，逐漸逝去。那是沒有數位相機的年代，網際網路也不發達，我到現在還是記不起來陳學長是怎麼計畫地帶我到處跑。所以這次我自己計畫，從機票、火車票、旅店及景點，都賴著網際網路的發達，在出發前就計畫好了，不過我到是花了我半個月的時間，在四處比價的情況下，才確定了行程，一個人旅行唯一的優點就是你的計畫你負責，不用擔心失敗的計畫拖累他人（謎之聲：言不由衷啊！）總的來說，這次唯一的敗筆就是行李太多了，可能是過度準備的結果！然而行李的因素，也考量在我交通移動計畫內，所以大體上我的計畫是http://yusung.blogspot.com/2009/08/blog-post.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-653630125944461080Tue, 04 Aug 2009 06:04:00 +00002009-08-04T02:24:24.509-04:00ConferenceRI have been to many conferences. Most of them are of Political Science. The common feature of these conferences is that the audience size is always small. If you get 10 audiences in your panel, your must be a big name.My presentation at the useR! 2009 was an exception. I had the biggest audience ever. I guess that there were at least over 60 people in the room. Had I known this, I should http://yusung.blogspot.com/2009/08/my-mi-presentation-at-user-2009.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-1644820149541563106Mon, 03 Aug 2009 02:08:00 +00002009-08-02T22:55:32.924-04:00ConferenceRSince my last attendance of the useR conference in Vienna, Austria, I have always wanted to go back. However, the high registration fee was always the factor that chills me off.This year, I am taking advantage of my last year as the student (the registration fee is cheaper) and the sponsorship of my two bosses, Andy and Jennifer, I went to the useR! 2009 conference. This year, the conference http://yusung.blogspot.com/2009/08/user-2009-conference-rennes-france.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-3671816065590761532Fri, 10 Jul 2009 08:50:00 +00002009-07-10T04:57:49.958-04:00ConferenceMiscThis is the last day of the useR conference. In the coffee break time, I am surfing on the web to plan my trip to Paris. It was 12 years ago since my last visit in France. The feeling is weird. And the keyboard of the computer lab in the Agrocampus Ouest University is also weird. The position of the keys is quite different than those of the US standard one. So far, this is the best conferencehttp://yusung.blogspot.com/2009/07/agrocampus-ouest-rennes-france.htmlnoreply@blogger.com (Yu-Sung Su)1tag:blogger.com,1999:blog-24082283.post-6762779145895637503Thu, 25 Jun 2009 03:40:00 +00002009-06-24T23:48:19.615-04:00RIn R, the way to delete a component in a list object is different from matrix and vector objects. For a vector, to delete an element:vec <- c(1, 2, 3)vec <- vec[-3]For a matrix, to delete a row or a column:mat <- matrix(c(1,2,3,4), 2, 2)mat2 <- mat[-1,] # delete a rowmat3 <- mat[,-1] # delete a columnFor a list, to delete a component:lis <- vector("list", 3)lis[[3]] <- NULLJust assign a NULL http://yusung.blogspot.com/2009/06/delete-list-component-in-r.htmlnoreply@blogger.com (Yu-Sung Su)0tag:blogger.com,1999:blog-24082283.post-2191476548770184775Sat, 20 Jun 2009 19:41:00 +00002009-06-20T16:16:13.274-04:00MiscSometimes, it is so weird how a voice from a radio program can suddenly catch your attention.After a regular weekly Friday Flushing dinner, Charlie and I went to a Korean supermarket to shop for grocery. I had nothing particular to buy. So I just wonder around the market. All of a sudden, I heard a sentense from a radio brocasting:"Don't make others your priorities, if you are just an option http://yusung.blogspot.com/2009/06/what-i-heard-from-radio-program.htmlnoreply@blogger.com (Yu-Sung Su)2tag:blogger.com,1999:blog-24082283.post-5027004597934282175Sun, 03 May 2009 18:51:00 +00002009-05-03T23:55:43.854-04:00Misc2009年4月30日下午4:15分，我開始辦理我的論文存放手續，5:00整，我走出學校的大樓，整整45分鐘，我在一連串奔跑中，正式完成我的博士求學生涯，接下來的就是畢業典禮了。回想我求學的過程，是順利吧，也充滿坎坷，順利地從幼稚園一路到博士畢業，中間從無間斷地，就這樣完成了所有學業，如果把近年來很流行的博士後研究算在內，老天真是不給我一個喘息的機會，讓我求學生涯中，硬是外加了這一段。去年12月，我就完成了那180頁的論文的口試，然而就因為這段博士後的插曲，為了工作簽證的關係，我硬是將那論文擺在家裡四個多月，而延遲了畢業，就這樣，我一點都沒有要畢業喜悅，而前天交上論文那一刻該爆發的喜悅，也因為四個月的「習慣」，而淡了。結束了，就像樂曲的finale...http://yusung.blogspot.com/2009/05/finale.htmlnoreply@blogger.com (Yu-Sung Su)4tag:blogger.com,1999:blog-24082283.post-5111920074924301555Sun, 22 Mar 2009 04:50:00 +00002009-03-23T11:06:28.212-04:00Misc這不是一則訴說紐約治安有多恐怖的網誌，而是奉勸人晚上了，能不要出門做事就不要！如果有人說我怕狗，我一定不同意，因為家裡也曾經養狗，所以對狗的熟悉度還是有的，不過我倒是很討厭大狗接近我。我印象中家裡曾養的二隻狗，一隻是柴犬，一隻是大土狗。那隻大土狗必須用大籠子關住牠，不然牠一出籠，那發足狂奔，口水四飛的樣子，著實令人不悅，我記得我總是透過家裡的玻璃門，看著牠放肆的樣子，也許吧，我對大狗的討厭是從牠而來。那隻柴犬呢，則是隻聰明的小帥狗，根據牠尾巴的捲度及翹度，我們都認為牠是隻聰明的公狗，不過呢，我在牠身上也學到了「狗是人最忠心的寵物」這句話是不成立的，牠第一次離家，就是為了牠的需求，一去不返，當初為了保有牠的自尊，沒想到...後來雖然奇蹟似的，半年後在嘉義某山區找到了牠，牠卻依然犯了天下狗兒都會犯的錯，再度離家不返，而這次，我們就再也沒見到牠了，陳小春一首「男人與公狗」似乎是不錯的類比。回到http://yusung.blogspot.com/2009/03/blog-post_22.htmlnoreply@blogger.com (Yu-Sung Su)2tag:blogger.com,1999:blog-24082283.post-3430610978250358285Sun, 08 Mar 2009 20:38:00 +00002009-03-21T16:29:17.117-04:00Misc我在紐約的時間，也許在倒數計時了...所以好多的第一次，也可能是最後一次了。昨天A君又帶團來了紐約，趁這個機會，我幫她約了一年前才聯絡上，在紐約某機關任職的K君，更巧的事，也因此聯絡上了一位，我一直聽說她在紐約但從來沒見過面的T君，於是我的紐約大學同學會就這樣開始了。這場聚會第一個爆點，就是T君的出現，她一身日系的打扮，一頭金髮，極瘦的身材，讓人覺得像是濱琦步出現在你面前，我定神看了一下，發現她右臂上有機械式的配件，宛如Iron Man的裝備，我那時心想，在大學時最有風格的一位女同學，現在仍是那麼勁爆！結果....那是她手的固定器，原來她一個月前划雪摔倒骨折，不過那付機械臂還真搭他的穿著，至於沒圖沒真相這件事，我就沒法子了，因為我們這群過三十的老同學，似乎沒有人有拍照的習慣。後來大家就去了Morton Steak House吃牛排，第二個爆點就來啦，A君、K君跟我都點了24 oz 的http://yusung.blogspot.com/2009/03/blog-post.htmlnoreply@blogger.com (Yu-Sung Su)2tag:blogger.com,1999:blog-24082283.post-3860068355018131499Tue, 03 Mar 2009 16:02:00 +00002009-03-03T11:09:14.991-05:00RStatisticsI got an email, asking me if our arm package can simulate tobit model to get simulated parameters. Indeed, arm does not suport tobit model. It only support sim() for lm, glm and mer classes in R. But it is not difficult to get a tobit verison of sim(). Here are the steps:1. fit a tobit (or double tobit) mode.library(AER)M1 <- tobit(y ~x) 2. get the coefficient vector.beta <- coef(M1) 3. http://yusung.blogspot.com/2009/03/simulate-parameters-of-tobit-model.htmlnoreply@blogger.com (Yu-Sung Su)1tag:blogger.com,1999:blog-24082283.post-6990055445998228190Mon, 09 Feb 2009 03:43:00 +00002009-02-08T22:43:30.936-05:00BUGSR I have got emails occasionally from JAGS users, asking about our new R package: R2jags. Basically, R2jags runs JAGS via R and makes postanalysis easier to be done in R. Taking advantage of the functions provided by JAGS, rjags and R2WinBUGS, R2jags allows users to run BUGS in the same way as they would do it in R2WinBUGS. Nonetheless, R2jags has some powerful features thathttp://yusung.blogspot.com/2009/02/our-new-r-package-r2jags.htmlnoreply@blogger.com (Yu-Sung Su)11tag:blogger.com,1999:blog-24082283.post-5032739721732238046Fri, 30 Jan 2009 20:06:00 +00002009-01-30T15:21:36.477-05:00RI did not know this and had to spend the whole week to figure out what's wrong with my code. I was updating the mi package. While I was doing code checking procedure, I always got a error message that I did not know what was going on. I remember last time when I encountered this, I just reduced the package dependency or simply don't use the S4 language. Anyway, here is the problem:mi is http://yusung.blogspot.com/2009/01/importing-generic-function-from-other.htmlnoreply@blogger.com (Yu-Sung Su)2tag:blogger.com,1999:blog-24082283.post-6128691711908547598Fri, 30 Jan 2009 01:57:00 +00002009-01-29T21:17:00.439-05:00BUGSRJust when I am about to convert myself into a Linux user, I found this. Running R in the command mode in the Window system is much faster. Here is a simple test:in the Window interface:> system.time(rnorm(1000000)%*%rnorm(1000000)) user system elapsed 1.06 0.05 1.14 in the command mode> system.time(rnorm(1000000)%*%rnorm(1000000)) user system elapsed 0.70 0.01 0.72 So ithttp://yusung.blogspot.com/2009/01/runing-r-in-command-mode-in-window.htmlnoreply@blogger.com (Yu-Sung Su)4tag:blogger.com,1999:blog-24082283.post-8269380468831594374Fri, 23 Jan 2009 21:00:00 +00002009-02-10T09:47:23.369-05:00BUGSLinuxRJAGS was developed to help those non-Window users to be able to use BUGS. However, I found it is very hard to install JAGS and its R package-rjags. Here is my note on how to install JAGS and rjags in Fedora 10. The note here should work for all Linux systems.Compile JAGS1. Download JAGS source code from JAGS website into any folder. The file you are looking for should look like this JAGS-1.0.3http://yusung.blogspot.com/2009/01/install-jags-and-rjags-in-fedora.htmlnoreply@blogger.com (Yu-Sung Su)25