Geographically Weighted Panel Regression – R

Deskripsi

Geographically Weighted Panel Regression didasarkan pada cabang statistik spasial. Dengan menggunakan bobot geografis, regresi panel berbobot geografis mencoba memecahkan residual dari pengelompokan regresi panel secara spasial. Untuk menyelidiki apakah residu mengelompok secara spasial, uji Moran’s I juga ditingkatkan. Selanjutnya, tiga uji statistik lokal disertakan untuk membantu pengguna memilih model.

Referensi utama adalah Fotheringham et al. (2003, ISBN:978-0-470-85525-6) dan Beenstock dan Felsen-stein (2019, ISBN:978-3-030-03614-0).

SOURCE CODE

library (lattice)
library (Matrix) 
library (sp)
library (spgwr)
library (maptools) 
xx<-readShapePoints ("D:/kemiskinan1.shp"[1]} 
хх
col.b.adapt<-
gwr.sel (Y1~X1+X2, data=xx, coords=cbind(xx$X, xx$Y), adapt=TRUE) adapt.gauss<- 
gwr (Y1~ X1+X2, data=xx, coords=cbind(xx$X, xx$Y), adapt=col.b.adapt, hatmatrix=TRUE)
names (adapt.gauss) h=adapt.gauss$bandwidth
h<-as.matrix(x)
b=h
for (i in 1:37) (b<-cbind(b,h)} 
b
je<-read.table("D:/je.txt")

w=function (b, je, i) {
b<-b[, i]
je<-je [, i]
k=je/b
a=k^2
c=(-0.5) *a
w-exp (c)
}

W1<- w (b, je,1)
W2<- w (b, je, 2)
W3<- w (b, je, 3)
W4<- w (b, je, 4)
W5<- w (b, je, 5)
W6<- w (b, je, 6)
W7<- w (b, je, 7)
W8<- w (b, je, 8)
W9<- w (b, je, 9)
W10<- w (b, je, 10)
W11<- w (b, je, 11)
W12<- w (b, je, 12)
W13<- w (b, je, 13)
W14<- w (b, je, 14)
W15<- w (b, je, 15)
W16<- w (b, je, 16)
W17<- w (b, je, 17)
W18<- w (b, je, 18)
W19<- w (b, je, 19)
W20<- w(b, je, 20)
W21<- w (b, je, 21) 

W22<- w (b, je, 22)
W23<- w (b, je, 23)
W24<- w (b, je, 24)
W25<- w (b, je, 25)
W26<- w (b, je, 26)
W27<- w (b, je, 27)
W28<- w (b, je, 28)
W29<- w (b, je, 29)
W30<- w (b, je, 30)
W11<- w (b, je, 31)
W32<- w (b, je, 32)
W33<- w (b, je, 33)
W34<- w (b, je, 34)
W35<- w (b, je, 35)
W136<- w (b, je, 36)
W37<- w (b, je, 37)
W38<- w (b, je, 38) 


W<-

rbind (W1, W2, W3, W4, W5, W6, W7, W8, W9, W10, W11, W12, W13, W14, W15, W16, W17, W18, W19, W20, W21, W22, W23, W24, W25, W26, W27, W28, W29, W30, W31, W32, W33, W34, W35, W36, W37, W38)

X<- read.table("D:/datax.txt")
Y<-read.table("D:/datay.txt")
olahdata = function (data, jumlahpengamatan, jumlahvariabel) {
data<-an.matrix(data) 
jumlahrow-nrow (data)
batas = jumlahpengamatan*jumlaltrow
X0= matrix (1, batas,1) 
data<-as.vector (data)
X<-matrix (data, batas, jumlahvariabel)
X<-cbind(X0,x)
}

X <- olahdata (X, 5, 2)
Y <- olahdata (Y, 5, 1)
Y < Y(, -1)
Y <- as.matrix(Y)

beta=function (X, Y, W, i) {
W<- W [, i ]
W<-as.matrix (W) 
jumlahrow=nrow (W)
pembobot=W 
for (i in 1:4) (pembobot <-rbind (pembobot, W)} pembobot
p1<-diag(190) 
diag (p1) <-pembobot 
p1

,beta37,beta38) 
Beta
tbeta=t (beta) 
tbeta
betaintersep =tbeta[, 1]
betaLPE=tbeta[, 2]
betaTPT=tbeta[, 3] 
min (betaintersep)
min (betaLPE)
min (betaTPT) 
max (betaintersep)
max (betaLPE)
max (betaTPT) 
rata0=sum (betaintersep)/38
rata0
rata1=sum (betaLPE) / 38 
rata1
rata2 =sum (betaTPT) / 38 
rata2
akarc=function (X, Y, W, i) f
W<-W [, i]
W<-as.matrix (W)
jumlahrow=nrow (W)
pembobot=w
for (i in 1:4) {pembobot<-rbind{pembobot, W}
pembobot 
p1<-diag (190)
diag (p1)<-pembobot
p1
W<-as.matrix (p1)
a = t(X) . W
b = solve (a . X)
C = b . a

CCt= c . t (C )
CCt
c = diag(CCt)
akarc = t(sqrt (c))
akarc = aa.matrix (akarc)
akarc
}

akarc1<-akarc (X, Y, W, 1)
akarc2<-akarc (X,Y, W, 2)
akarc3<-akarc (X, Y, W, 3)
akarc4<-akarc (X, Y,W, 4)
akarc5<-akarc (X, Y, W, 5)
akarc6<-akare (X, Y, W, 6)
akare?<-akarc (X, Y, W, 7)
akarc8<-akarc (X, Y, W, 8)
akare9<-akare (X, Y, W, 9)
akarc10<-akarc (X, Y, W, 10)
akarcii<-akarc (X, Y, W, 11)
akarc12<-akarc (X, Y, W, 12)
akarc13<-akarc (X, Y, W, 13)
akarc14<-akarc (X, Y, W, 14)
akarc15<-akarc (X, Y, W, 15)
akarc16<-akarc (X, Y, W, 16)
akarc17<-akarc (X, Y, W, 17)
akare18<-akarc (X, Y, W, 18)
akarc19<-akarc (X, Y, W, 19)
akarc20<-akarc (X, Y, W, 20)
akarc21<-akarc (X, Y, W, 21)
akarc22<-akarc (X,Y, W, 22)
akarc23<-akarc (X, Y, W, 23)
akare24<-akarc (X, Y, W, 24)
akare25<-akarc (X, Y, W, 25)
akarc26<-akarc (X, Y, W, 26)
akare27<-akarc (X, Y,W,27)
akarc28<-akarc (X, Y,W,28)
akarc29<-akarc (X, Y, W, 29)
akarc30<-akarc (X, Y, W, 30)
akare31<-akarc (X, Y, W, 31)
akarc32<-akarc (X, Y, W, 32)
akarc33<-akarc (X, Y, W, 33)
akarc34<-akarc (X, Y, W, 34)
akarc35<-akarc (X, Y, W, 35)
akarc36<-akarc (X, Y, W, 36)
akarc37<-akarc (X, Y, W, 37)
akarc38<-akarc (X, Y, W, 38))

akarc<- आ rbind (akarc1, akarc2, akarc3, akarc4, akarc5, akarc6, akarc7, akarc8, akarc9, akare10, akarc11, akarc12, akarc13, akarc14, akarc15, akarc16, akarc17, akarc18, akarc19, akarc20, akarc21, akarc22, akarc23, akarc24, akarc25, akarc26, akarc27, akarc28, akarc29, akarc30, akarc31 ,akarc32, akarc33, akarc34, akare35, akarc36, akarc37, akarc38) akarc

m1=function (X, Y, W, 1, j). {
w<-w[,1]
W<-as.matrix (W)
jumlahrow=nrow (W)
pembobot=W 
for (i in 1:4) (pembobot<-rbind (pembobot, W) }
pembobot
pl<-diag (190)
diag (pl) <-pembobot
pl

W<-as.matrix (pl) 
a = t (X)  * W 
b = solve (a . X )
c = b . a
Xt=t (x)

L = Xt [, j] . C 
}

L1=ml (X, Y, W, 1, 1)
L2=ml (X, Y, W, 2, 2)
L3=ml (X, Y, W, 3,3)
L4=ml (X, Y, W, 4, 4)
L5=ml (X, Y, W, 5, 5)
L6=ml (X, Y, W, 6, 6)
L7=ml (X, Y, W, 7, 7)
L8=ml (X, Y, W, 8, 8)
L9=ml (X, Y, W, 9, 9)
L10=ml (X, Y, W, 10, 10)
L11=ml (X, Y, W, 11,11)
L12=mi (X,Y, W, 12, 12)
L13=ml (X, Y, W,13,13)
L14=ml (X, Y, W, 14, 14)
L15=ml (X,Y, W, 15, 15)
L16=ml (X, Y, W, 16,16)
L17=ml (X, Y, W,17,17)
L18=ml (X, Y, W,18,18)
L19=ml (X, Y, W,19,19)
L20=ml (X, Y, W, 20, 201
L21=ml (X, X, W, 21, 21)
L22=ml (X, Y, W, 22, 22)
L23=ml (X, Y, W, 23, 23)
L24=ml (X, Y, W, 24, 24)
L25=ml (X, Y, W, 25, 25)
L26=ml (X, Y, W, 26, 26)
L27=ml (X, Y, W, 27,27)
L28=ml (X, Y, W, 28,28)
L29=ml (X, Y, W, 29, 29)
L30=ml (X, Y, W, 30, 30)
L31=ml (X, Y, W, 31,31)
L32=ml (X, Y, W, 32, 32)
L33=ml (X, Y, W, 33, 33)
L34=ml (X, Y ,W, 34, 34)
L35=ml (X,Y, W, 35, 35)
L36=ml (X,Y, W, 36, 36)
L37=ml (X, Y, W, 37,37)
L38=ml (X, Y, W, 38, 38)
L39=ml (X, Y, W, 39, 39)
L40=ml (X, Y, W, 40,40)
L41=m1 (X, Y, W, 41, 41)
L42=ml (X, Y, W, 42,42)
L43=ml (X, Y, W, 43, 43)
L44=ml (X, Y, W, 44, 44)
L45=ml (X, Y ,W, 45, 45)
L46=ml (X, Y, W, 46, 46) 


L108, L109, L110, L111, L112, L113, L114, L115, L116, L117, L118, L119, L120, L121, L173, L174, L175, L176, L177, L178, L179, L180, L181, L182, L183, L184, L185, L186, L187
L

ci = array (1, dim = c (190,1)) 
I = diag (190)
diag (1) <-ci
I
f = I - L
g = t (f) %*% f
 g <-as.matrix(g)
si1 = sum (diag(g))
si1
j = g^2
si2 = sum (diag(j))
si2
u = t (x)%*%X
V = solve(u)
H = X%*%V%*%t(x)
V - I - H
JKGH1 - t (Y)%*%V%*%Y
JKGH1
sigmatopi = sqrt (JKGH1/si1) 
sigmatopi

sigmatopi - rbind (sigmatopi,sigmatopi,signatopi)
sigmatopi
tsigmatopi = t(sigmatopi)
ts = tsigmatopi
for (i in 1:37) (ts< - rbind(ts, tsigmatopi))
ts

salahbaku = ts akarc
salahbaku
t = tbeta/salahbaku
t

ratay = sum (Y)/190

tabeta = tbeta
for (i in 1:4) (tabeta<-rbind(tabeta, tbeta))

Ytopi = tabeta*X 
Ytopi
Y1< - Ytopi (, 1)
Y1< - as.matrix(Y1)

Y2<-Ytopi (, 2)
Y2<-as.matrix (Y2)
Y3<-Ytopi (,3)
13<-as.matrix (Y3)
Ytopi=Y1+Y2+Y3
Ytopi

11 = (Y- Ytopi)^2
111 = 11
for (i in 1:189) (111<- cbind(111, 111))
W = as.matrix (W)
p=w
for (i in 1:4) (p<-cbind(p, W))
w=p
for (i in 1:4) (<-rbind (w,p))
kl=sum (W*111)

jil = ( Y - ratay)^2
j i l = j l
for (1 in 1:189) (jil<-rbind(j i l, j l) }
hl=sum (W*jil)

Rkuadrat=kl/hl 
Rkuadrat