[程式] Hansen2000門檻模型與R

看板Statistics作者supa666w (灰色哲學)時間14年前 (2011/05/11 01:18)推噓2(2推 0噓 21→)

留言23則, 3人參與討論串1/1

------------------------------------------------------------------------ [軟體程式類別]:R [程式問題]:程式碼 [軟體熟悉度]: 新手(不到1個月) [問題敘述]: 小弟利用R來跑Hansen2000年發表的論文裡面的資料 data與code都是作者公佈的他在code裡面有設定可在同質變數或是異質變數的假設下跑回歸 code分別是h <-1(假設異質) 或h <-0(假設同質) 在h<-1的情況下我可以跑出完整的報表可是當我輸入h <-0再輸入回歸式時就出現錯誤在b1 - ser1 * z : 非調和陣列可是我只是在h那裡改0或1而已其他都設定好了也沒變動不知道哪位前輩可以為小弟解惑? 第一次在板上po文無理之處煩請見諒 [程式範例]: # Load procedures and data # source("C:/r/thr_het.R") source("c:/r/thr_est.R") data <- read.table("c:/r/dur_john.dat") # Exclude missing variables and oil states # k <- ncol(data) indx <- as.matrix(1-(data[,5]== -999))%*%matrix(c(1),1,k) data <- as.matrix(data[indx>0]) data <- matrix(data,nrow=nrow(data)/k,ncol=k) indx <- as.matrix(1-(data[,6]== -999))%*%matrix(c(1),1,k) data <- as.matrix(data[indx>0]) data <- matrix(data,nrow=nrow(data)/k,ncol=k) indx <- as.matrix(1-(data[,10]== -999))%*%matrix(c(1),1,k) data <- as.matrix(data[indx>0]) data <- matrix(data,nrow=nrow(data)/k,ncol=k) indx <- as.matrix(1-(data[,11]== -999))%*%matrix(c(1),1,k) data <- as.matrix(data[indx>0]) data <- matrix(data,nrow=nrow(data)/k,ncol=k) indx <- as.matrix(data[,2]== 1)%*%matrix(c(1),1,k) data <- as.matrix(data[indx>0]) data <- matrix(data,nrow=nrow(data)/k,ncol=k) # Make data transformations # diff <- log(data[,6])-log(data[,5]) q <- data[,5] gdp60 <- log(q) iony <- log(data[,9]/100) pgro <- log(data[,8]/100+.05) sch <- log((data[,10])/100) lit <- data[,11] dat <- cbind(diff,gdp60,iony,pgro,sch,q,lit) # Program switches # rep <- 1000 h <- 1 ------------->問題所在 na <- rbind("GNP_Gwth","GDP_1960","Inv/GDP","Pop_Gwth","School","GDP_1960", "Literacy") dum <- rbind(2,3,4,5) # Estimate First Sample Split, Using Output as Threshold # qhat1 <- thr_est(dat,na,1,dum,6,h) --------------------------------------------------------------------------- -- ※ 發信站: 批踢踢實業坊(ptt.cc) ◆ From: 123.194.244.102

→

gsuper

05/11 01:30, , 1^F

05/11 01:30, 1^F

→

gsuper

05/11 01:30, , 2^F

05/11 01:30, 2^F

→

gsuper

05/11 01:31, , 3^F

05/11 01:31, 3^F