different prompt

.gitignore

@@ -200,3 +200,4 @@ data/mit_long-term/ECGs.csv
 data/mit_long-term/ECGs.csv
 .DS_Store
 docs_wip/readme/README_popularity.py
+paper/OHBM2022/OHBM2022_Makowski.mp4

docs_wip/conf.py

@@ -93,6 +93,10 @@ def find_version():
 napoleon_use_rtype = False
 add_module_names = False  # If true, the current module name will be prepended to all description
 
+# -- Options for ipython directive  ----------------------------------------
+
+ipython_promptin = ">"  # "In [%d]:"
+ipython_promptout = ">"  # "Out [%d]:"
 
 # -- Options for myst_nb ---------------------------------------------------
 nb_execution_mode = "force"

neurokit2/complexity/complexity.py

@@ -32,8 +32,8 @@
 def complexity(signal, which=["fast", "medium"], delay=1, dimension=2, tolerance="sd", **kwargs):
     """**Complexity and Chaos Analysis**
 
-    Measuring complexity refers to the quantification of various concepts, such as **chaos, entropy,
-    unpredictability, and fractal dimension**.
+    Measuring the complexity of a signal refers to the quantification of various aspects related to
+    concepts such as **chaos**, **entropy**, **unpredictability**, and **fractal dimension**.
 
     .. tip::
 
@@ -50,11 +50,11 @@ def complexity(signal, which=["fast", "medium"], delay=1, dimension=2, tolerance
     separately, to gain more control over the parameters and information that you get.
 
     The categorization by "computation time" is based on our preliminary `benchmarking study
-    <https://neurokit2.readthedocs.io/en/latest/studies/complexity_benchmark.html>`_ results:
+    <https://neuropsychology.github.io/NeuroKit/studies/complexity_benchmark.html>`_ results:
 
     .. figure:: ../../studies/complexity_benchmark/figures/computation_time-1.png
        :alt: Complexity Benchmark (Makowski).
-       :target: https://neurokit2.readthedocs.io/en/latest/studies/complexity_benchmark.html
+       :target: https://neuropsychology.github.io/NeuroKit/studies/complexity_benchmark.html
 
     Parameters
     ----------
@@ -68,11 +68,11 @@ def complexity(signal, which=["fast", "medium"], delay=1, dimension=2, tolerance
         See :func:`complexity_delay` to estimate the optimal value for this parameter.
     dimension : int
         Embedding Dimension (*m*, sometimes referred to as *d* or *order*). See
-        :func:`complexity_dimension()` to estimate the optimal value for this parameter.
+        :func:`complexity_dimension` to estimate the optimal value for this parameter.
     tolerance : float
         Tolerance (often denoted as *r*), distance to consider two data points as similar. If
         ``"sd"`` (default), will be set to :math:`0.2 * SD_{signal}`. See
-        :func:`complexity_tolerance()` to estimate the optimal value for this parameter.
+        :func:`complexity_tolerance` to estimate the optimal value for this parameter.
 
     Returns
     --------
@@ -83,8 +83,7 @@ def complexity(signal, which=["fast", "medium"], delay=1, dimension=2, tolerance
 
     See Also
     --------
-    entropy_permutation, entropy_differential, entropy_svd, fractal_katz, fractal_petrosian,
-    fractal_sevcik, fisher_information, complexity_hjorth, complexity_rqa
+    complexity_delay, complexity_dimension, complexity_tolerance
 
     Examples
     ----------

studies/complexity_benchmark/README.Rmd

@@ -56,7 +56,7 @@ The script to generate the data can be found at ...
 
 Generated 5 types of signals, to which we added different types of noise.
 
-```{r message=FALSE, warning=FALSE, fig.height=20, fig.width=12}
+```{r message=FALSE, warning=FALSE, fig.height=20, fig.width=12, cache=TRUE}
 library(tidyverse)
 library(easystats)
 library(patchwork)
@@ -268,7 +268,7 @@ colors <- c(
 
 ### Computation Time
 
-```{r computation_time, message=FALSE, warning=FALSE, fig.width=16*1.25, fig.height=10*1.25}
+```{r computation_time, message=FALSE, warning=FALSE, fig.width=16*1.25, fig.height=10*1.25, cache=TRUE}
 order <- df |>
   group_by(Index) |>
   summarize(Duration = median(Duration)) |>
@@ -392,7 +392,7 @@ estimate_relation(model) |>
 ### Correlation
 
 
-```{r message=FALSE, warning=FALSE, fig.width=16, fig.height=15}
+```{r message=FALSE, warning=FALSE, fig.width=16, fig.height=15, cache=TRUE}
 data <- df |>
   mutate(i = paste(Signal, Length, Noise_Type, Noise_Intensity, sep = "__")) |>
   select(i, Index, Result) |>
@@ -461,7 +461,7 @@ cor <- get_cor(data)
   - Remove **RR**  because it's slower.
 
 
-```{r message=FALSE, warning=FALSE, include=FALSE, eval=FALSE}
+```{r message=FALSE, warning=FALSE, include=FALSE, eval=FALSE, cache=TRUE}
 cor |>
   cor_lower() |>
   filter(Parameter1 %in% names(data), Parameter2 %in% names(data)) |>
@@ -487,21 +487,10 @@ filter(averagetime, Index %in% c("FuzzyEn", "FuzzyApEn"))
 filter(averagetime, Index %in% c("SVDEn", "FuzzycApEn"))
 filter(averagetime, Index %in% c("CPEn", "CRPEn"))
 filter(averagetime, Index %in% c("NLDFD", "RR"))
+```
 
 
-# NLFD | RR
-# NLFD | RR
-# - Drop RR because it's slower
-# H (uncorrected) | H (corrected)
-# - ??
-# SVDEn | FuzzyEn
-# - Drop FuzzyEn because it's slower
-
-# Hasselman positively correlated with most of the others
-# - RR: much slower
-
-
-
+```{r message=FALSE, warning=FALSE, cache=TRUE}
 data <- data |>
   select(
     -`CREn (B)`,
@@ -517,12 +506,15 @@ data <- data |>
     -FuzzycApEn,
     -CPEn,
     -RR,
+    -MFDFA_HDelta,
+    -FuzzyRCMSEn,
+    -`CREn (1000)`, -`CREn (100)`,
+    -RQA_VEn, -RQA_LEn
   )
 
 cor <- get_cor(data)
 ```
 
-
 <!-- ### Hierarchical CLustering -->
 
 
@@ -544,7 +536,7 @@ cor <- get_cor(data)
 ### Factor Analysis
 
 
-```{r message=FALSE, warning=FALSE, include=FALSE, eval=FALSE}
+```{r message=FALSE, warning=FALSE, fig.width=12, fig.height=18}
 r <- correlation::cor_smooth(as.matrix(cor))
 
 plot(parameters::n_factors(data, cor = r))

studies/complexity_benchmark/README.md

@@ -246,6 +246,33 @@ cor <- get_cor(data)
 -   **NLDFD**, and **RR**
     -   Remove **RR** because it’s slower.
 
+``` r
+data <- data |>
+  select(
+    -`CREn (B)`,
+    -`CREn (D)`, -`ShanEn (D)`,
+    -`CREn (r)`, -`ShanEn (r)`,
+    # -`CREn (C)`, -`ShanEn (C)`,
+    -`PSDFD (Voss1998)`,
+    -`RangeEn (A)`, -`RangeEn (Ac)`,
+    -FI,
+    -MMSEn,
+    -`H (corrected)`,
+    -FuzzyApEn,
+    -FuzzycApEn,
+    -CPEn,
+    -RR,
+    -MFDFA_HDelta,
+    -FuzzyRCMSEn,
+    -`CREn (1000)`, -`CREn (100)`,
+    -RQA_VEn, -RQA_LEn
+  )
+
+cor <- get_cor(data)
+```
+
+![](../../studies/complexity_benchmark/figures/unnamed-chunk-9-1.png)<!-- -->
+
 <!-- ### Hierarchical CLustering -->
 <!-- ```{r message=FALSE, warning=FALSE} -->
 <!-- n <- parameters::n_clusters(as.data.frame(t(data)), standardize = FALSE) -->
@@ -262,4 +289,80 @@ cor <- get_cor(data)
 
 ### Factor Analysis
 
+``` r
+r <- correlation::cor_smooth(as.matrix(cor))
+
+plot(parameters::n_factors(data, cor = r))
+```
+
+![](../../studies/complexity_benchmark/figures/unnamed-chunk-10-1.png)<!-- -->
+
+``` r
+rez <- parameters::factor_analysis(data, cor = r, n = 14, rotation = "varimax", sort = TRUE, fm="wls")
+# rez <- parameters::principal_components(data, n = 5, sort = TRUE)
+# rez
+
+col <- gsub('[[:digit:]]+', '', names(rez)[2])
+closest <- colnames(select(rez, starts_with(col)))[apply(select(rez, starts_with(col)), 1, \(x) which.max(abs(x)))]
+
+loadings <- attributes(rez)$loadings_long |>
+  mutate(
+    Loading = abs(Loading),
+    Component = fct_relevel(Component, rev(names(select(rez, starts_with(col))))),
+    Variable = fct_rev(fct_relevel(Variable, rez$Variable))
+  )
+
+colors <- setNames(see::palette_material("rainbow")(length(levels(loadings$Component))), levels(loadings$Component))
+
+
+p1 <- loadings |>
+  # filter(Variable == "CD") |>
+  ggplot(aes(x = Variable, y = Loading)) +
+  geom_bar(aes(fill = Component), stat = "identity") +
+  geom_vline(xintercept = c("SD", "Length", "Noise", "Random"), color = "red") +
+  geom_vline(xintercept = head(cumsum(sort(table(closest))), -1) + 0.5) +
+  scale_y_continuous(expand = c(0, 0)) +
+  scale_fill_material_d("rainbow") +
+  coord_flip() +
+  theme_minimal() +
+  guides(fill = guide_legend(reverse = TRUE)) +
+  labs(x = NULL) +
+  theme(
+    axis.text.y = element_text(
+      color = rev(colors[closest]),
+      face = rev(ifelse(rez$Variable %in% c("SD", "Length", "Noise", "Random"), "italic", "plain")),
+      hjust = 0.5
+    ),
+    axis.text.x = element_blank(),
+    plot.title = element_text(hjust = 0.5),
+    panel.grid.major = element_blank(),
+    panel.grid.minor = element_blank()
+  )
+
+p2 <- order |>
+  mutate(Duration = 1 + Duration * 10000) |>
+  filter(Index %in% loadings$Variable) |>
+  mutate(Index = fct_relevel(Index, levels(loadings$Variable))) |>
+  ggplot(aes(x = log10(Duration), y = Index)) +
+  geom_bar(aes(fill = log10(Duration)), stat = "identity") +
+  geom_hline(yintercept = head(cumsum(sort(table(closest))), -1) + 0.5) +
+  scale_x_reverse(expand = c(0, 0)) +
+  # scale_x_log10(breaks = 10**seq(0, 4), labels = function(x) sprintf("%g", x), expand=c(0, 0)) +
+  scale_y_discrete(position = "right") +
+  scale_fill_viridis_c(guide = "none") +
+  labs(x = "Computation Time", y = NULL) +
+  theme_minimal() +
+  theme(
+    axis.text.y = element_blank(),
+    axis.text.x = element_blank(),
+    plot.title = element_text(hjust = 0.5),
+    panel.grid.major = element_blank(),
+    panel.grid.minor = element_blank()
+  )
+
+(p2 | p1) + patchwork::plot_annotation(title = "Computation Time and Factor Loading", theme = theme(plot.title = element_text(hjust = 0.5, face = "bold")))
+```
+
+![](../../studies/complexity_benchmark/figures/unnamed-chunk-10-2.png)<!-- -->
+
 ## References