generated from RealityBending/TemplateResults
-
Notifications
You must be signed in to change notification settings - Fork 1
/
6_Summary.Rmd
250 lines (201 loc) · 11.1 KB
/
6_Summary.Rmd
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
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
---
output: html_document
editor_options:
chunk_output_type: console
---
### Metaclustering Matrix
```{r, message=FALSE, warning=FALSE}
table <- table_clusters # [c("Index", "EGA", "EFA_3", "Mixture")]
# Initialize matrix
m <- matrix(data=NA, nrow=nrow(table), ncol=nrow(table), dimnames = list(table$Index, table$Index))
for(row in row.names(m)) {
for(col in colnames(m)) {
if(row == col) {
m[row, col] <- 0
next
}
subset <- table[table$Index %in% c(row, col), ]
rez <- sapply(subset[2:ncol(subset)], function(x) {
if(any(is.na(x))) {
NA
} else {
ifelse(length(unique(x[!is.na(x)])) == 1, 0, 1)
}
})
m[row, col] <- sum(rez, na.rm = TRUE) / length(na.omit(rez))
}
}
```
```{r, message=FALSE, warning=FALSE}
m <- correlation::cor_sort(m, distance = "raw", hclust_method = "average")
# colnames(m) <- clean_names(colnames(m))
# rownames(m) <- clean_names(rownames(m))
# hrv_cols <- clean_names(hrv_cols)
# colnames(z) <- clean_names(colnames(z))
abs(m - 1) %>%
as.data.frame() %>%
rownames_to_column("Index1") %>%
datawizard::reshape_longer(cols = hrv_cols, colnames_to = "Index2", values_to = "Probability") %>%
mutate(Index1 = factor(Index1, levels = row.names(m)),
Index2 = factor(Index2, levels = rev(row.names(m)))) %>%
ggplot(aes(x = Index1, y = Index2)) +
geom_tile(aes(fill = Probability)) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
labs(x = NULL, y = NULL) +
scale_fill_gradient2(low = "#2196F3", mid = "#9C27B0", high = "#FF5722", midpoint = 0.5) +
labs(title = "Metaclustering Heatmap", subtitle = "Probability of being grouped together accross multiple methods.")
# ggsave("figures/figure_probability_matrix.png", width=12*0.9, height=9*0.9, dpi = 450)
```
### Hierarchical Structure of Metaclustering
```{r, message=FALSE, warning=FALSE}
model <- hclust(as.dist(m), method="average")
clusters <- cutree(model, h = 0.8)
clusters
# clusters2 <- c(
# pNN50 = 1, pNN20 = 1, CVI = 1, RMSSD = 1, S = 1, HTI = 1, MCVNN = 1,
# MadNN = 1, IQRNN = 1, CVNN = 1, SDNN = 1, SD2 = 1,
# MeanNN = 2, MedianNN = 2,
# ApEn = 2, ShanEn = 2, MSE = 2,
# DFA_alpha2_DimMean = 3,
# LF = 3, VHF = 3, HF = 3, LnHF = 3,
# CMSE = 3, RCMSE = 3,
# LZC = 3, SampEn = 3,
# FuzzyEn = 3, TINN = 3,
# DFA_alpha1_ExpMean = 3, CD = 3, DFA_alpha1_ExpRange = 3, KFD = 3,
# SD1SD2 = 4, HFn = 4, HFD = 4,
# DFA_alpha1_DimMean = 4, DFA_alpha2_ExpRange = 4, DFA_alpha2_DimRange = 4,
# DFA_alpha2 = 4, DFA_alpha2_ExpMean = 4, CSI_Modified = 4,
# CSI = 4, LFn = 4, LFHF = 4, DFA_alpha1 = 4,
# PAS = 5, PSS = 5, PIP = 5, IALS = 5,
# C1a = 6, Ca = 6, PI = 6, C2a = 6,
# DFA_alpha1_DimRange = 6,
# AI = 6, GI = 6, SI = 6)
# identify center of each cluster
dis <- parameters::cluster_centers(datawizard::data_transpose(z), clusters)
# reverse dis to have larger value for more central params
centrality <- datawizard::data_rescale(attributes(dis)$distance, to = c(max(attributes(dis)$distance), min(attributes(dis)$distance)))
names(centrality) <- hrv_cols
plot_dendrogram(model, clusters = clusters, ylim = -0.2, nudge_y = -0.03, angle = 90, points = TRUE, dis = centrality, h=0.8) +
# Lines to show "larger" clusters
# geom_line(data = data.frame(y = c(-0.02, -0.02), x = c(0, 5)), aes(x = x, y = y), size = 0.5, color = "#607D8B") +
# geom_line(data = data.frame(y = c(-0.02, -0.02), x = c(6, 15)), aes(x = x, y = y), size = 0.5, color = "#607D8B") +
# geom_line(data = data.frame(y = c(-0.02, -0.02), x = c(16, 22)), aes(x = x, y = y), size = 0.5, color = "#607D8B") +
# geom_line(data = data.frame(y = c(-0.02, -0.02), x = c(23, 34)), aes(x = x, y = y), size = 0.5, color = "#607D8B") +
# geom_line(data = data.frame(y = c(-0.02, -0.02), x = c(35, 42)), aes(x = x, y = y), size = 0.5, color = "#607D8B") +
theme(legend.position = "none")
# ggsave(file = "figures/meta-dendrogram.svg", plot=dendro, width=15, height=10)
```
### Metaclustering Network
#### Centrality
```{r, message=FALSE, warning=FALSE}
nodes <- data.frame(name = row.names(m),
Cluster = as.character(clusters))
row.names(nodes) <- NULL
edges <- m %>%
as.data.frame() %>%
datawizard::data_rename_rows(NULL) %>%
datawizard::data_rename(NULL) %>%
datawizard::reshape_longer(rows_to = "from",
names_to = "to",
values_to = "Probability") %>%
mutate(to = as.numeric(to)) %>%
filter(Probability > 0.8)
graph <- tidygraph::tbl_graph(nodes = nodes, edges = edges, directed = FALSE) %>%
tidygraph::activate(nodes) %>%
mutate(Centrality = tidygraph::centrality_degree())
as.list(graph)$nodes %>%
group_by(Cluster) %>%
arrange(desc(Centrality)) %>%
knitr::kable()
```
#### Network Graph
```{r, message=FALSE, warning=FALSE}
graph %>%
ggraph::ggraph(layout = "graphopt") +
geom_edge_arc(aes(colour = Probability, width = Probability), strength = 0.05) +
geom_node_point(aes(colour = Cluster, size = Centrality)) +
geom_node_text(aes(label = name), fontface = "bold", repel = FALSE) +
ggraph::theme_graph() +
scale_edge_width(range = c(0.05, 0.5), guide = "none") +
scale_color_material_d(palette = "rainbow", guide = "none") +
scale_edge_color_gradientn(limits = c(0, 1), colours=c("#607D8B", "#263238"), guide = "none") +
scale_size_continuous(range = c(4, 20), guide = "none")
```
<!-- #### Create table -->
<!-- ```{r, message=FALSE, warning=FALSE, eval=FALSE, echo=FALSE} -->
<!-- level_1 <- c(rep("Distribution", 17), -->
<!-- rep("Frequency/Complexity", 28), -->
<!-- rep("Harmony", 12)) -->
<!-- level_2 <- c(rep("Dispersion", 12), -->
<!-- rep("Centrality", 5), -->
<!-- rep("Absolute Frequency/Complexity", 15), -->
<!-- rep("Relative Frequency/Complexity", 13), -->
<!-- rep("Heart Rate Fragmentation", 4), -->
<!-- rep("Heart Rate Asymmetry", 8)) -->
<!-- indices <- c("pNN50", "pNN20", "CVI", "RMSSD", "S", "HTI", "MCVNN", -->
<!-- "MadNN", "IQRNN", "CVNN", "SDNN", "SD2", -->
<!-- "MeanNN", "MedianNN", -->
<!-- "ApEn", "ShanEn", "MSE", -->
<!-- "DFA α2 DimMean", -->
<!-- "LF", "VHF", "HF", "LnHF", -->
<!-- "CMSE", "RCMSE", "LZC", "SampEn", "FuzzyEn", "TINN", "DFA α1 ExpMean", -->
<!-- "CD", "DFA α1 ExpRange", "KFD", -->
<!-- "SD1SD2", "HFn", "HFD", -->
<!-- "DFA α1 DimMean", "DFA α2 ExpRange", "DFA α2 DimRange", -->
<!-- "DFA α2", "DFA α2 ExpMean", "CSI (modified)", -->
<!-- "LFn", "CSI", "LFHF", "DFA α1", -->
<!-- "PAS", "IALS", "PSS", "PIP", -->
<!-- "C1a", "C2a", "Ca", "PI", -->
<!-- "DFA α1 DimRange", -->
<!-- "AI", "GI", "SI") -->
<!-- labels <- c("Proportion of successive NN interval differences larger than 50ms", -->
<!-- "Proportion of successive NN interval differences larger than 20ms", -->
<!-- "Cardiac Vagal Index", -->
<!-- "Root mean square of successive NN interval differences", -->
<!-- "Area of ellipse in Poincaré plot", -->
<!-- "Integral of the density of the NN interval histogram divided by its height", -->
<!-- "MadNN divided by MedianNN", -->
<!-- "The median absolute deviation of the NN intervals", -->
<!-- "The interquartile range (IQR) of the NN intervals", -->
<!-- "SDNN divided by MeanNN", "The standard deviation of the RR intervals", -->
<!-- "The spread of NN intervals on the Poincaré plot along the line of identity.", -->
<!-- "The mean of the NN intervals.", "The median of the NN intervals.", -->
<!-- "The Approximate Entropy", "The Shannon Entropy", "The Multiscale Entropy", -->
<!-- "The multifractal detrended fluctuation analysis (MDFA) corresponding to long-term correlations. Dimmean is the mean of singularity dimensions", -->
<!-- "Power spectrum in the frequency range of 0.04-0.15 Hz", -->
<!-- "Power spectrum in the frequency range of 0.4-0.5 Hz", -->
<!-- "Power spectrum in the frequency range of 0.15-0.4 Hz", -->
<!-- "The natural logarithm of HF", -->
<!-- "The Composite Multiscale Entropy", "The Refined Composite Multiscale Entropy", -->
<!-- "The Lempel-Ziv complexity", "The Sample Entropy", "The Fuzzy Entropy", -->
<!-- "The baseline width of the NN interval histogram", -->
<!-- "The MDFA corresponding to short-term correlation. ExpMean is the mean of singularity exponents", -->
<!-- "Correlation Dimension", "The MDFA corresponding to short-term correlation. ExpRange is the range of singularity exponents", -->
<!-- "Katz Fractal Dimension", -->
<!-- "The ratio between short and long term fluctuations of the NN intervals", "The normalized HF", "Higuchi Fractal Dimension", -->
<!-- "The MDFA corresponding to short-term correlation. DimMean is the mean of singularity dimensions", "The MDFA corresponding to long-term correlation. ExpRange is the range of singularity exponents", "The MDFA corresponding to long-term correlation. DimRange is the range of singularity dimensions", -->
<!-- "The DFA corresponding to long-term correlation", "The MDFA corresponding to long-term correlation. ExpMean is the mean of singularity exponents", "The Cardiac Symapathetic Index (modified)", -->
<!-- "The normalized LF", "The Cardiac Sympathetic Index", "The ratio between LF and HF", "The DFA corresponding to short-term corrleation", -->
<!-- "The percentage of NN intervals in alternation segments", -->
<!-- "The inverse of the average length of the acceleration/deceleration segments", -->
<!-- "The percentage of short segments", -->
<!-- "The percentage of inflection points of the NN intervals series", -->
<!-- "The contributions of heart rate accelerations to short-term HRV", -->
<!-- "The contributions of heart rate accelerations to long-term HRV", -->
<!-- "The total contributions of heart rate accelerations to HRV", -->
<!-- "The Porta's Index", -->
<!-- "The MDFA corresponding to short-term correlation. DimRange is the range of singularity dimensions", -->
<!-- "The Area Index", "The Guzik's Index", "The Slope Index") -->
<!-- df_summary <- data.frame(level_1, level_2, indices, labels) -->
<!-- df_summary$Centrality <- as.data.frame(centrality)[df_summary$indices, ] -->
<!-- df_summary <- df_summary %>% -->
<!-- group_by(level_1, level_2) %>% -->
<!-- arrange(desc(Centrality), .by_group = TRUE) %>% -->
<!-- mutate(Centrality = round(Centrality, digits = 1)) -->
<!-- # rename("Level 1" = "level_1", -->
<!-- # "Level 2" = "level_2", -->
<!-- # "HRV Indices" = "indices", -->
<!-- # "Description" = "labels") -->
<!-- write.csv(df_summary, "figures/Summary_Table.csv", row.names = FALSE) -->
<!-- ``` -->