From c30b3ef56a04dc7b00ef6e311dd9366806ba73f8 Mon Sep 17 00:00:00 2001
From: Matthew Simmons <matthewsimmons@192-168-1-101.tpgi.com.au>
Date: Sun, 20 Sep 2020 15:06:06 +1000
Subject: [PATCH] Add a 'none' option to not apply normalization, in case
 another method outside the function is desired.

---
 DESCRIPTION            |  2 +-
 R/find_HDoutliers.R    | 12 +++++++-----
 man/find_HDoutliers.Rd |  6 +++---
 3 files changed, 11 insertions(+), 9 deletions(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index d5aaf43..00bdaad 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -27,4 +27,4 @@ BugReports: https://github.com/pridiltal/stray/issues
 License: GPL-2
 Encoding: UTF-8
 LazyData: true
-RoxygenNote: 7.1.0
+RoxygenNote: 7.1.1
diff --git a/R/find_HDoutliers.R b/R/find_HDoutliers.R
index 6b02d2f..c5f4c15 100644
--- a/R/find_HDoutliers.R
+++ b/R/find_HDoutliers.R
@@ -9,8 +9,8 @@
 #' @param k Number of neighbours considered.
 #' @param knnsearchtype A character vector indicating the search type for k- nearest-neighbors.
 #' @param normalize Method to normalize the columns of the data. This prevents variables with large variances
-#'  having disproportional influence on Euclidean distances. Two options are available "standardize" or "unitize".
-#'  Default is set to "unitize"
+#'  having disproportional influence on Euclidean distances. Three options are available: "standardize", "unitize" or "none".
+#'  Default is set to "unitize". If "none" is passed, no normalization will be used (not recommended.)
 #' @param p Proportion of possible candidates for outliers. This defines the starting point for the
 #' bottom up searching algorithm. Default is set to 0.5.
 #' @param tn Sample size to calculate an emperical threshold. Default is set to 50.
@@ -38,7 +38,8 @@
 #'
 find_HDoutliers <- function(data, alpha = 0.01, k = 10,
                             knnsearchtype = "brute",
-                            normalize = "unitize", p = 0.5, tn =50) {
+                            normalize = c("unitize", "standardize", "none"), p = 0.5, tn =50) {
+  normalize <- match.arg(normalize, choices = c("unitize", "standardize", "none"))
   data <- as.matrix(data)
   r <- nrow(data)
   data[is.infinite(data)] <- NA
@@ -62,8 +63,9 @@ find_HDoutliers <- function(data, alpha = 0.01, k = 10,
   standardize <- function(z) {
     (z - stats::median(z)) / stats::IQR(z)
   }
-
-  data <- apply(as.matrix(naomit_data), 2, normalize)
+  if (normalize != "none") {
+    data <- apply(as.matrix(naomit_data), 2, normalize)
+  }
   out <- use_KNN(data, alpha, k = k, knnsearchtype = knnsearchtype, p = p, tn=tn)
   outliers <- tag[out$outliers]
   type <- as.factor(ifelse(1:r %in% outliers,
diff --git a/man/find_HDoutliers.Rd b/man/find_HDoutliers.Rd
index 600d0b0..6f3353d 100644
--- a/man/find_HDoutliers.Rd
+++ b/man/find_HDoutliers.Rd
@@ -9,7 +9,7 @@ find_HDoutliers(
   alpha = 0.01,
   k = 10,
   knnsearchtype = "brute",
-  normalize = "unitize",
+  normalize = c("unitize", "standardize", "none"),
   p = 0.5,
   tn = 50
 )
@@ -26,8 +26,8 @@ distances between exemplars.}
 \item{knnsearchtype}{A character vector indicating the search type for k- nearest-neighbors.}
 
 \item{normalize}{Method to normalize the columns of the data. This prevents variables with large variances
-having disproportional influence on Euclidean distances. Two options are available "standardize" or "unitize".
-Default is set to "unitize"}
+having disproportional influence on Euclidean distances. Three options are available: "standardize", "unitize" or "none.
+Default is set to "unitize". If "none" is passed, no normalization will be used (not recommended.)}
 
 \item{p}{Proportion of possible candidates for outliers. This defines the starting point for the
 bottom up searching algorithm. Default is set to 0.5.}