From fa7c546b507bf601da2f232f583660654a3d935e Mon Sep 17 00:00:00 2001
From: huizezhang-sherry <huizezhangsh@gmail.com>
Date: Thu, 23 May 2024 23:44:59 -0500
Subject: [PATCH] add more useful indexes

---
 DESCRIPTION                           |  6 ++-
 NAMESPACE                             |  5 ++
 R/interesting-indices.r               | 67 +++++++++++++++++++++++++--
 _pkgdown.yml                          |  2 +
 man/dcor.Rd                           | 17 +++++++
 man/dcor2d.Rd                         | 13 ------
 man/guided_tour.Rd                    |  6 +--
 man/indexes.Rd                        | 15 ++++++
 man/search_jellyfish.Rd               |  4 +-
 man/{splines2d.Rd => spline-loess.Rd} |  7 ++-
 man/stringy.Rd                        | 11 +++++
 11 files changed, 127 insertions(+), 26 deletions(-)
 create mode 100644 man/dcor.Rd
 delete mode 100644 man/dcor2d.Rd
 create mode 100644 man/indexes.Rd
 rename man/{splines2d.Rd => spline-loess.Rd} (72%)
 create mode 100644 man/stringy.Rd

diff --git a/DESCRIPTION b/DESCRIPTION
index d28ebb9f..ba51f602 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -41,11 +41,13 @@ Suggests:
     rmarkdown,
     tidyr,
     covr,
-    plotly
+    plotly,
+    cassowaryr,
+    minerva
 License: MIT + file LICENSE
 LazyData: true
 URL: https://github.com/ggobi/tourr
 BugReports: https://github.com/ggobi/tourr/issues
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.1
 Encoding: UTF-8
 VignetteBuilder: knitr
diff --git a/NAMESPACE b/NAMESPACE
index 47466dc0..bb145d3f 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -12,6 +12,8 @@ S3method(plot,path_index)
 S3method(print,history_array)
 S3method(print,tour_path)
 S3method(str,history_array)
+export(MIC)
+export(TIC)
 export(anchored_orthogonal_distance)
 export(andrews)
 export(angular_breaks)
@@ -40,6 +42,7 @@ export(center)
 export(cmass)
 export(cumulative_radial)
 export(dcor2d)
+export(dcor2d_2)
 export(dependence_tour)
 export(display_andrews)
 export(display_density2d)
@@ -75,6 +78,7 @@ export(lda_pp)
 export(linear_breaks)
 export(little_tour)
 export(local_tour)
+export(loess2d)
 export(manual_slice)
 export(mapColors)
 export(mapShapes)
@@ -109,6 +113,7 @@ export(search_posse)
 export(slice_index)
 export(sphere_data)
 export(splines2d)
+export(stringy)
 export(thaw)
 importFrom(grDevices,dev.cur)
 importFrom(grDevices,dev.flush)
diff --git a/R/interesting-indices.r b/R/interesting-indices.r
index 49f6d234..462f0314 100644
--- a/R/interesting-indices.r
+++ b/R/interesting-indices.r
@@ -1,11 +1,45 @@
+#' Scagnostic indexes.
+#'
+#' Compute the scagnostic measures from the cassowaryr package
+#' @export
+stringy <- function(){
+  function(mat){
+    cassowaryr::sc_stringy(mat[,1], mat[,2])
+  }
+}
+
+
+#' Maximum and total information coefficient index.
+#'
+#' Compute the maximum and total information coefficient indexes,
+#' see \code{minerva::mine}.
+#'
+#' @rdname indexes
+#' @export
+MIC <- function(){
+  function(mat){
+    minerva::mine(mat[,1], mat[,2], alpha = 0.3,  est = "mic_e")$MIC
+  }
+}
+
+#' @rdname indexes
+#' @export
+TIC <- function(){
+  function(mat){
+    minerva::mine(mat[,1], mat[,2], est = "mic_e", alpha = 0.3)$TIC
+  }
+}
+
 #' Distance correlation index.
 #'
-#' Computes the distance correlation based index on
-#' 2D projections of the data.
+#' Computes the distance correlation based index on 2D projections of the data.
+#' \code{dcor2d_2} uses the faster implementation of the distance correlation
+#' for bivariate data, see \code{energy::dcor2d}.
 #'
 #' @keywords hplot
 #' @importFrom stats na.omit
 #' @export
+#' @rdname dcor
 dcor2d <- function() {
   function(mat) {
     xy <- na.omit(data.frame(x = mat[, 1], y = mat[, 2]))
@@ -14,15 +48,26 @@ dcor2d <- function() {
   }
 }
 
-#' Spline based index.
+#' @rdname dcor
+#' @export
+dcor2d_2 <- function() {
+  function(mat) {
+    xy <- na.omit(data.frame(x = mat[, 1], y = mat[, 2]))
+    measure <- with(xy, energy::dcor2d(x, y, type = "U"))
+    return(measure)
+  }
+}
+
+#' Spline/loess based index.
 #'
 #' Compares the variance in residuals of a fitted
-#' spline model to the overall variance to find
+#' spline/loess model to the overall variance to find
 #' functional dependence in 2D projections
 #' of the data.
 #'
 #' @keywords hplot
 #' @importFrom stats residuals var
+#' @rdname spline-loess
 #' @export
 splines2d <- function() {
   function(mat) {
@@ -38,6 +83,20 @@ splines2d <- function() {
   }
 }
 
+#' @rdname spline-loess
+#' @export
+loess2d <- function() {
+  function(mat) {
+    mat <- as.data.frame(mat)
+    colnames(mat) <- c("x", "y")
+    loess_fit <- loess(y ~ x, data = mat, span = 0.05)
+    loess_fit2 <- loess(x ~ y, data = mat, span = 0.05)
+    measure <- max(1 - var(residuals(loess_fit), na.rm = T) / var(mat$y, na.rm = T),
+                   1 - var(residuals(loess_fit2), na.rm = T) / var(mat$y, na.rm = T)
+    )
+    return(measure)
+  }
+}
 
 
 #' Normality index.
diff --git a/_pkgdown.yml b/_pkgdown.yml
index eb873103..ee29c591 100644
--- a/_pkgdown.yml
+++ b/_pkgdown.yml
@@ -52,6 +52,8 @@ reference:
   - slice_index
   - norm_bin
   - norm_kol
+  - stringy
+  - MIC
 - title: Search functions
   desc: >
     Functions for index optimisation in guided tour
diff --git a/man/dcor.Rd b/man/dcor.Rd
new file mode 100644
index 00000000..6f9f9669
--- /dev/null
+++ b/man/dcor.Rd
@@ -0,0 +1,17 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/interesting-indices.r
+\name{dcor2d}
+\alias{dcor2d}
+\alias{dcor2d_2}
+\title{Distance correlation index.}
+\usage{
+dcor2d()
+
+dcor2d_2()
+}
+\description{
+Computes the distance correlation based index on 2D projections of the data.
+\code{dcor2d_2} uses the faster implementation of the distance correlation
+for bivariate data, see \code{energy::dcor2d}.
+}
+\keyword{hplot}
diff --git a/man/dcor2d.Rd b/man/dcor2d.Rd
deleted file mode 100644
index 9c5e68d4..00000000
--- a/man/dcor2d.Rd
+++ /dev/null
@@ -1,13 +0,0 @@
-% Generated by roxygen2: do not edit by hand
-% Please edit documentation in R/interesting-indices.r
-\name{dcor2d}
-\alias{dcor2d}
-\title{Distance correlation index.}
-\usage{
-dcor2d()
-}
-\description{
-Computes the distance correlation based index on
-2D projections of the data.
-}
-\keyword{hplot}
diff --git a/man/guided_tour.Rd b/man/guided_tour.Rd
index ba99114b..83375437 100644
--- a/man/guided_tour.Rd
+++ b/man/guided_tour.Rd
@@ -7,13 +7,13 @@
 guided_tour(
   index_f,
   d = 2,
-  alpha = 0.5,
   cooling = 0.99,
   max.tries = 25,
   max.i = Inf,
   search_f = search_geodesic,
   n_jellies = 30,
   n_sample = 100,
+  alpha = 0.5,
   ...
 )
 }
@@ -22,8 +22,6 @@ guided_tour(
 
 \item{d}{target dimensionality}
 
-\item{alpha}{the initial size of the search window, in radians}
-
 \item{cooling}{the amount the size of the search window should be adjusted
 by after each step}
 
@@ -37,6 +35,8 @@ a better projection before giving up}
 
 \item{n_sample}{number of samples to generate if \code{search_f} is \code{\link{search_polish}}}
 
+\item{alpha}{the initial size of the search window, in radians}
+
 \item{...}{arguments sent to the search_f}
 }
 \description{
diff --git a/man/indexes.Rd b/man/indexes.Rd
new file mode 100644
index 00000000..3549b6bc
--- /dev/null
+++ b/man/indexes.Rd
@@ -0,0 +1,15 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/interesting-indices.r
+\name{MIC}
+\alias{MIC}
+\alias{TIC}
+\title{Maximum and total information coefficient index.}
+\usage{
+MIC()
+
+TIC()
+}
+\description{
+Compute the maximum and total information coefficient indexes,
+see \code{minerva::mine}.
+}
diff --git a/man/search_jellyfish.Rd b/man/search_jellyfish.Rd
index 68037ebf..c4fec8d1 100644
--- a/man/search_jellyfish.Rd
+++ b/man/search_jellyfish.Rd
@@ -4,7 +4,7 @@
 \alias{search_jellyfish}
 \title{An jellyfish optimisers for the projection pursuit guided tour}
 \usage{
-search_jellyfish(current, index, tries, max.tries = Inf, min.tries = 30, ...)
+search_jellyfish(current, index, tries, max.tries = 50, ...)
 }
 \arguments{
 \item{current}{starting projection, a list of basis of class "multi-bases"}
@@ -13,7 +13,7 @@ search_jellyfish(current, index, tries, max.tries = Inf, min.tries = 30, ...)
 
 \item{tries}{the counter of the outer loop of the opotimiser}
 
-\item{max.tries, min.tries}{the maximum/minimum number of iteration before giving up}
+\item{max.tries}{the maximum number of iteration before giving up}
 
 \item{...}{other arguments being passed into the \code{search_jellyfish()}}
 }
diff --git a/man/splines2d.Rd b/man/spline-loess.Rd
similarity index 72%
rename from man/splines2d.Rd
rename to man/spline-loess.Rd
index 05cf368e..760ce288 100644
--- a/man/splines2d.Rd
+++ b/man/spline-loess.Rd
@@ -2,13 +2,16 @@
 % Please edit documentation in R/interesting-indices.r
 \name{splines2d}
 \alias{splines2d}
-\title{Spline based index.}
+\alias{loess2d}
+\title{Spline/loess based index.}
 \usage{
 splines2d()
+
+loess2d()
 }
 \description{
 Compares the variance in residuals of a fitted
-spline model to the overall variance to find
+spline/loess model to the overall variance to find
 functional dependence in 2D projections
 of the data.
 }
diff --git a/man/stringy.Rd b/man/stringy.Rd
new file mode 100644
index 00000000..fbfe7313
--- /dev/null
+++ b/man/stringy.Rd
@@ -0,0 +1,11 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/interesting-indices.r
+\name{stringy}
+\alias{stringy}
+\title{Scagnostic indexes.}
+\usage{
+stringy()
+}
+\description{
+Compute the scagnostic measures from the cassowaryr package
+}