-
Notifications
You must be signed in to change notification settings - Fork 0
/
error_table_script.r
174 lines (149 loc) · 9.22 KB
/
error_table_script.r
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
# Remove any previous objects in the workspace
rm(list=ls(all=TRUE))
graphics.off()
source("functions_NCAvNLME_2016.r")
library(plyr)
library(ggplot2)
library(reshape2)
library(stringr)
# Choose the directory you wish to compile results from
master.dir <- "E:/hscpw-df1/Data1/Jim Hughes/2016"
setwd(master.dir)
file.name <- "collated_error_data.csv"
dir.r16 <- "01_MDV"
df.r16 <- read.csv(paste(dir.r16, file.name, sep="/"))
dir.r78 <- "04_SS"
df.r78 <- read.csv(paste(dir.r78, file.name, sep="/"))
dir.r910 <- "06_genKA"
df.r910 <- read.csv(paste(dir.r910, file.name, sep="/"))
dir.r1114 <- "07_noBOV"
df.r1114 <- read.csv(paste(dir.r1114, file.name, sep="/"))
df.all <- rbind(df.r16, df.r78, df.r910, df.r1114)
df.all$Scen <- df.all$SCEN
df.all$Scen[df.all$RUN %% 2 == 0] <- df.all$SCEN[df.all$RUN %% 2 == 0] + 9
df.all$Run <- ceiling(df.all$RUN/2)
meth.str1 <- str_extract(df.all$Model, "Run[0-9]_Scen[0-9]{2}")
meth.str2 <- str_extract(df.all$Model, "Run[0-9]{2}_Scen[0-9]{2}")
meth.str <- c(meth.str1[!is.na(meth.str1)], meth.str2[!is.na(meth.str2)])
df.all$Method <- as.numeric(substr(meth.str, nchar(meth.str), nchar(meth.str)))
df.m1 <- df.all[df.all$Method == 1,]
df.m3 <- df.all[df.all$Method == 3,]
search.termstat <- as.character(unique(df.all$TermStat))
search.covcode <- as.character(unique(df.all$CovCode))
search.errcode <- as.character(unique(df.all$ErrCode))
search.termdesc <- as.character(unique(df.all$TermDesc))
per.onelevel <- function(x, col, search) {
unlist(llply(1:length(search), function(i, x, col, search) {
length(x[col][x[col] == search[i]])/dim(x)[1]*100
}, x = x, col = col, search = search))
}
per.twolevel <- function(x, col, s.grid) {
unlist(llply(1:dim(s.grid)[1], function(i, x, col, search) {
length(x[col[1]][x[col[1]] == search[i,1] &
x[col[2]] == search[i,2]])/dim(x)[1]*100
}, x = x, col = col, search = s.grid))
}
#Percentages of successful vs. unsuccessful minimisation
term.all.all <- per.onelevel(df.all, "TermStat", search.termstat)
term.all.m1 <- per.onelevel(df.m1, "TermStat", search.termstat)
term.all.m3 <- per.onelevel(df.m3, "TermStat", search.termstat)
#Run
term.run.all <- ddply(df.all, .(Run), function(x) per.onelevel(x, "TermStat", search.termstat))
term.run.m1 <- ddply(df.m1, .(Run), function(x) per.onelevel(x, "TermStat", search.termstat))
term.run.m3 <- ddply(df.m3, .(Run), function(x) per.onelevel(x, "TermStat", search.termstat))
#Scen
term.scen.all <- ddply(df.all, .(Scen), function(x) per.onelevel(x, "TermStat", search.termstat))
term.scen.m1 <- ddply(df.m1, .(Scen), function(x) per.onelevel(x, "TermStat", search.termstat))
term.scen.m3 <- ddply(df.m3, .(Scen), function(x) per.onelevel(x, "TermStat", search.termstat))
#Run+Scen
term.runscen.all <- ddply(df.all, .(Run, Scen), function(x) per.onelevel(x, "TermStat", search.termstat))
term.runscen.m1 <- ddply(df.m1, .(Run, Scen), function(x) per.onelevel(x, "TermStat", search.termstat))
term.runscen.m3 <- ddply(df.m3, .(Run, Scen), function(x) per.onelevel(x, "TermStat", search.termstat))
#Percentages of successful vs. unsuccessful covariance step
cov.all.all <- per.onelevel(df.all, "CovCode", search.covcode)
cov.all.m1 <- per.onelevel(df.m1, "CovCode", search.covcode)
cov.all.m3 <- per.onelevel(df.m3, "CovCode", search.covcode)
#Run
cov.run.all <- ddply(df.all, .(Run), function(x) per.onelevel(x, "CovCode", search.covcode))
cov.run.m1 <- ddply(df.m1, .(Run), function(x) per.onelevel(x, "CovCode", search.covcode))
cov.run.m3 <- ddply(df.m3, .(Run), function(x) per.onelevel(x, "CovCode", search.covcode))
#Scen
cov.scen.all <- ddply(df.all, .(Scen), function(x) per.onelevel(x, "CovCode", search.covcode))
cov.scen.m1 <- ddply(df.m1, .(Scen), function(x) per.onelevel(x, "CovCode", search.covcode))
cov.scen.m3 <- ddply(df.m3, .(Scen), function(x) per.onelevel(x, "CovCode", search.covcode))
#Run+Scen
cov.runscen.all <- ddply(df.all, .(Run, Scen), function(x) per.onelevel(x, "CovCode", search.covcode))
cov.runscen.m1 <- ddply(df.m1, .(Run, Scen), function(x) per.onelevel(x, "CovCode", search.covcode))
cov.runscen.m3 <- ddply(df.m3, .(Run, Scen), function(x) per.onelevel(x, "CovCode", search.covcode))
#Percentages of different errors - "None", "Rounding errors", "Obj close to infinity", "Zero gradient", "Reached max evaluations"
err.all.all <- per.onelevel(df.all, "ErrCode", search.errcode)
err.all.m1 <- per.onelevel(df.m1, "ErrCode", search.errcode)
err.all.m3 <- per.onelevel(df.m3, "ErrCode", search.errcode)
#Run
err.run.all <- ddply(df.all, .(Run), function(x) per.onelevel(x, "ErrCode", search.errcode))
err.run.m1 <- ddply(df.m1, .(Run), function(x) per.onelevel(x, "ErrCode", search.errcode))
err.run.m3 <- ddply(df.m3, .(Run), function(x) per.onelevel(x, "ErrCode", search.errcode))
#Scen
err.scen.all <- ddply(df.all, .(Scen), function(x) per.onelevel(x, "ErrCode", search.errcode))
err.scen.m1 <- ddply(df.m1, .(Scen), function(x) per.onelevel(x, "ErrCode", search.errcode))
err.scen.m3 <- ddply(df.m3, .(Scen), function(x) per.onelevel(x, "ErrCode", search.errcode))
#Run+Scen
err.runscen.all <- ddply(df.all, .(Run, Scen), function(x) per.onelevel(x, "ErrCode", search.errcode))
err.runscen.m1 <- ddply(df.m1, .(Run, Scen), function(x) per.onelevel(x, "ErrCode", search.errcode))
err.runscen.m3 <- ddply(df.m3, .(Run, Scen), function(x) per.onelevel(x, "ErrCode", search.errcode))
#Percentages of different descriptions - "None", "Parameter estimate near boundary",
#"R matrix algorithmically singular", "Problems occurred during minimisation",
#"S matrix unobtainable", "Due to last iteration",
#"Problems with individual", "Due to next iteration"
desc.all.all <- per.onelevel(df.all, "TermDesc", search.termdesc)
desc.all.m1 <- per.onelevel(df.m1, "TermDesc", search.termdesc)
desc.all.m3 <- per.onelevel(df.m3, "TermDesc", search.termdesc)
#Run
desc.run.all <- ddply(df.all, .(Run), function(x) per.onelevel(x, "TermDesc", search.termdesc))
desc.run.m1 <- ddply(df.m1, .(Run), function(x) per.onelevel(x, "TermDesc", search.termdesc))
desc.run.m3 <- ddply(df.m3, .(Run), function(x) per.onelevel(x, "TermDesc", search.termdesc))
#Scen
desc.scen.all <- ddply(df.all, .(Scen), function(x) per.onelevel(x, "TermDesc", search.termdesc))
desc.scen.m1 <- ddply(df.m1, .(Scen), function(x) per.onelevel(x, "TermDesc", search.termdesc))
desc.scen.m3 <- ddply(df.m3, .(Scen), function(x) per.onelevel(x, "TermDesc", search.termdesc))
#Run+Scen
desc.runscen.all <- ddply(df.all, .(Run, Scen), function(x) per.onelevel(x, "TermDesc", search.termdesc))
desc.runscen.m1 <- ddply(df.m1, .(Run, Scen), function(x) per.onelevel(x, "TermDesc", search.termdesc))
desc.runscen.m3 <- ddply(df.m3, .(Run, Scen), function(x) per.onelevel(x, "TermDesc", search.termdesc))
qplot(factor(TermStat), data = df.all, geom = "bar", facets = ~Run)
qplot(factor(CovCode), data = df.all, geom = "bar", facets = ~Run)
qplot(factor(ErrCode), data = df.all, geom = "bar", facets = ~Run)
qplot(factor(TermDesc), data = df.all, geom = "bar", facets = ~Run)
qplot(factor(TermStat), data = df.all, geom = "bar", facets = ~Method)
qplot(factor(CovCode), data = df.all, geom = "bar", facets = ~Method)
qplot(factor(ErrCode), data = df.all, geom = "bar", facets = ~Method)
qplot(factor(TermDesc), data = df.all, geom = "bar", facets = ~Method)
err.all.m1[2]/term.all.m1[2]
err.all.m3[2]/term.all.m3[2]
err.run.m1[3]/term.run.m1[3]
err.run.m3[3]/term.run.m3[3]
err.scen.m1[3]/term.scen.m1[3]
err.scen.m3[3]/term.scen.m3[3]
err.runscen.m1[4]/term.runscen.m1[4]
err.runscen.m3[4]/term.runscen.m3[4]
desc.all.m1[2]/(cov.all.m1[2] - term.all.m1[2])*100
desc.all.m3[2]/(cov.all.m3[2] - term.all.m3[2])*100
desc.all.m1[3]/(cov.all.m1[2] - term.all.m1[2])*100
desc.all.m3[3]/(cov.all.m3[2] - term.all.m3[2])*100
#Determine search terms for
s.grid.errdesc.all <- expand.grid(search.covcode, search.termdesc, stringsAsFactors = F)
err.desc.all.all <- per.twolevel(df.all, c("CovCode","TermDesc"), s.grid.errdesc.all)
s.grid.errdesc <- s.grid.errdesc.all[err.desc.all.all != 0, ]
cols.errdesc <- paste(s.grid.errdesc$Var1, s.grid.errdesc$Var2)
#Run
errdesc.run.all <- ddply(df.all, .(Run), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
errdesc.run.m1 <- ddply(df.m1, .(Run), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
errdesc.run.m3 <- ddply(df.m3, .(Run), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
#Scen
errdesc.scen.all <- ddply(df.all, .(Scen), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
errdesc.scen.m1 <- ddply(df.m1, .(Scen), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
errdesc.scen.m3 <- ddply(df.m3, .(Scen), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
#Run+Scen
errdesc.runscen.all <- ddply(df.all, .(Run, Scen), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
errdesc.runscen.m1 <- ddply(df.m1, .(Run, Scen), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))
errdesc.runscen.m3 <- ddply(df.m3, .(Run, Scen), function(x) per.twolevel(x, c("ErrCode","TermDesc"), s.grid.errdesc))