1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
|
"pcaiv" <- function (dudi, df, scannf = TRUE, nf = 2) {
lm.pcaiv <- function(x, df, weights, use) {
if (!inherits(df, "data.frame"))
stop("data.frame expected")
reponse.generic <- x
begin <- "reponse.generic ~ "
fmla <- as.formula(paste(begin, paste(names(df), collapse = "+")))
df <- cbind.data.frame(reponse.generic, df)
lm0 <- lm(fmla, data = df, weights = weights)
if (use == 0)
return(predict(lm0))
else if (use == 1)
return(residuals(lm0))
else if (use == -1)
return(lm0)
else stop("Non convenient use")
}
if (!inherits(dudi, "dudi"))
stop("dudi is not a 'dudi' object")
df <- data.frame(df)
if (!inherits(df, "data.frame"))
stop("df is not a 'data.frame'")
if (nrow(df) != length(dudi$lw))
stop("Non convenient dimensions")
weights <- dudi$lw
isfactor <- unlist(lapply(as.list(df), is.factor))
for (i in 1:ncol(df)) {
if (!isfactor[i])
df[, i] <- scalewt(df[, i], weights)
}
tab <- data.frame(apply(dudi$tab, 2, lm.pcaiv, df = df, use = 0,
weights = dudi$lw))
X <- as.dudi(tab, dudi$cw, dudi$lw, scannf = scannf, nf = nf,
call = match.call(), type = "pcaiv")
X$X <- df
X$Y <- dudi$tab
U <- as.matrix(X$c1) * unlist(X$cw)
U <- as.matrix(dudi$tab) %*% U
U <- data.frame(U)
row.names(U) <- row.names(dudi$tab)
names(U) <- names(X$li)
X$ls <- U
U <- as.matrix(X$c1) * unlist(X$cw)
U <- data.frame(t(as.matrix(dudi$c1)) %*% U)
row.names(U) <- names(dudi$li)
names(U) <- names(X$li)
X$as <- U
w <- apply(X$ls, 2, function(x) coefficients(lm.pcaiv(x,
df, weights, -1)))
w <- data.frame(w)
names(w) <- names(X$l1)
X$fa <- w
fmla <- as.formula(paste("~ ", paste(names(df), collapse = "+")))
w <- scalewt(model.matrix(fmla, data = df)[,-1], weights) * weights
w <- t(w) %*% as.matrix(X$l1)
w <- data.frame(w)
X$cor <- w
return(X)
}
"plot.pcaiv" <- function (x, xax = 1, yax = 2, ...) {
if (!inherits(x, "pcaiv"))
stop("Use only with 'pcaiv' objects")
if (x$nf == 1) {
warnings("One axis only : not yet implemented")
return(invisible())
}
if (xax > x$nf)
stop("Non convenient xax")
if (yax > x$nf)
stop("Non convenient yax")
def.par <- par(no.readonly = TRUE)
on.exit(par(def.par))
layout(matrix(c(1, 2, 3, 4, 4, 5, 4, 4, 6), 3, 3),
respect = TRUE)
par(mar = c(0.1, 0.1, 0.1, 0.1))
# modif mail P. Giraudoux 25/10/2004
s.arrow(na.omit(x$fa), xax, yax, sub = "Loadings", csub = 2,
clabel = 1.25)
s.arrow(na.omit(x$cor), xax = xax, yax = yax, sub = "Correlation",
csub = 2, clabel = 1.25)
s.corcircle(x$as, xax, yax, sub = "Inertia axes", csub = 2)
s.match(x$li, x$ls, xax, yax, clabel = 1.5, sub = "Scores and predictions",
csub = 2)
if (inherits(x, "cca"))
s.label(x$co, xax, yax, clabel = 0, cpoint = 3, add.plot = TRUE)
if (inherits(x, "cca"))
s.label(x$co, xax, yax, clabel = 1.25, sub = "Species",
csub = 2)
else s.arrow(x$c1, xax = xax, yax = yax, sub = "Variables",
csub = 2, clabel = 1.25)
scatterutil.eigen(x$eig, wsel = c(xax, yax))
}
"print.pcaiv" <- function (x, ...) {
if (!inherits(x, "pcaiv"))
stop("to be used with 'pcaiv' object")
cat("Principal Component Analysis with Instrumental Variables\n")
cat("call: ")
print(x$call)
cat("class: ")
cat(class(x), "\n")
cat("\n$rank (rank) :", x$rank)
cat("\n$nf (axis saved) :", x$nf)
cat("\n\neigen values: ")
l0 <- length(x$eig)
cat(signif(x$eig, 4)[1:(min(5, l0))])
if (l0 > 5)
cat(" ...\n\n")
else cat("\n\n")
sumry <- array("", c(3, 4), list(rep("", 3), c("vector",
"length", "mode", "content")))
sumry[1, ] <- c("$eig", length(x$eig), mode(x$eig), "eigen values")
sumry[2, ] <- c("$lw", length(x$lw), mode(x$lw), "row weigths (from dudi)")
sumry[3, ] <- c("$cw", length(x$cw), mode(x$cw), "col weigths (from dudi)")
print(sumry, quote = FALSE)
cat("\n")
sumry <- array("", c(3, 4), list(rep("", 3), c("data.frame",
"nrow", "ncol", "content")))
sumry[1, ] <- c("$Y", nrow(x$Y), ncol(x$Y), "Dependant variables")
sumry[2, ] <- c("$X", nrow(x$X), ncol(x$X), "Explanatory variables")
sumry[3, ] <- c("$tab", nrow(x$tab), ncol(x$tab), "modified array (projected variables)")
print(sumry, quote = FALSE)
cat("\n")
sumry <- array("", c(4, 4), list(rep("", 4), c("data.frame",
"nrow", "ncol", "content")))
sumry[1, ] <- c("$c1", nrow(x$c1), ncol(x$c1), "PPA Pseudo Principal Axes")
sumry[2, ] <- c("$as", nrow(x$as), ncol(x$as), "Principal axis of dudi$tab on PAP")
sumry[3, ] <- c("$ls", nrow(x$ls), ncol(x$ls), "projection of lines of dudi$tab on PPA")
sumry[4, ] <- c("$li", nrow(x$li), ncol(x$li), "$ls predicted by X")
print(sumry, quote = FALSE)
cat("\n")
sumry <- array("", c(4, 4), list(rep("", 4), c("data.frame",
"nrow", "ncol", "content")))
sumry[1, ] <- c("$fa", nrow(x$fa), ncol(x$fa), "Loadings (CPC as linear combinations of X")
sumry[2, ] <- c("$l1", nrow(x$l1), ncol(x$l1), "CPC Constraint Principal Components")
sumry[3, ] <- c("$co", nrow(x$co), ncol(x$co), "inner product CPC - Y")
sumry[4, ] <- c("$cor", nrow(x$cor), ncol(x$cor), "correlation CPC - X")
print(sumry, quote = FALSE)
cat("\n")
}
summary.pcaiv <- function(object, ...){
thetitle <- "Principal component analysis with instrumental variables"
cat(thetitle)
cat("\n\n")
NextMethod()
appel <- as.list(object$call)
dudi <- eval.parent(appel$dudi)
cat(paste("Total unconstrained inertia (",deparse(appel$dudi),"): ", sep = ""))
cat(signif(sum(dudi$eig), 4))
cat("\n\n")
cat(paste("Inertia of" ,deparse(appel$dudi),"explained by", deparse(appel$df), "(%): "))
cat(signif(sum(object$eig) / sum(dudi$eig) * 100, 4))
cat("\n\n")
cat("Decomposition per axis:\n")
sumry <- array(0, c(object$nf, 7), list(1:object$nf, c("iner", "inercum", "inerC", "inercumC", "ratio", "R2", "lambda")))
sumry[, 1] <- dudi$eig[1:object$nf]
sumry[, 2] <- cumsum(dudi$eig[1:object$nf])
varpro <- apply(object$ls, 2, function(x) sum(x * x * object$lw))
sumry[, 3] <- varpro
sumry[, 4] <- cumsum(varpro)
sumry[, 5] <- cumsum(varpro)/cumsum(dudi$eig[1:object$nf])
sumry[, 6] <- object$eig[1:object$nf]/varpro
sumry[, 7] <- object$eig[1:object$nf]
print(sumry, digits = 3)
invisible(sumry)
}
|
