predict for DSC now honors the formula in the DSC. Updated man pages.

DESCRIPTION

@@ -13,12 +13,12 @@ Authors@R: c(person("Michael", "Hahsler", role = c("aut", "cre", "cph"),
             person("Dalibor", "Krleža", role = c("ctb"))
             )
 Description: A framework for data stream modeling and associated data mining tasks such as clustering and classification. The development of this package was supported in part by NSF IIS-0948893, NSF CMMI 1728612, and NIH R21HG005912. Hahsler et al (2017) <doi:10.18637/jss.v076.i14>.
-Depends: R (>= 3.5.0), methods, proxy (>= 0.4-7), magrittr
-Imports: clue, cluster, clusterGeneration, dbscan (>= 1.0-0), fpc, graphics, grDevices, MASS, mlbench, rpart, Rcpp (>= 0.11.4), stats, utils
+Depends: R (>= 3.5.0), methods, magrittr
+Imports: clue, cluster, clusterGeneration, dbscan (>= 1.0-0), fpc, graphics, grDevices, MASS, mlbench, rpart, Rcpp (>= 0.11.4), stats, utils, proxy (>= 0.4-7)
 Suggests: animation, DBI, dplyr, rJava, RSQLite, testthat, knitr
 URL: https://github.com/mhahsler/stream
 BugReports: https://github.com/mhahsler/stream/issues
 LinkingTo: Rcpp, BH
 License: GPL-3
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.1
 Roxygen: list(markdown = TRUE)

NAMESPACE

@@ -108,6 +108,7 @@ S3method(update,DSF)
 S3method(update,DSF_Convolve)
 S3method(update,DSF_Downsample)
 S3method(update,DSF_ExponentialMA)
+S3method(update,DSF_FeatureSelection)
 S3method(update,DSF_Func)
 S3method(update,DSF_Scale)
 S3method(update,DSF_dplyr)
@@ -164,6 +165,7 @@ export(DSFP)
 export(DSF_Convolve)
 export(DSF_Downsample)
 export(DSF_ExponentialMA)
+export(DSF_FeatureSelection)
 export(DSF_Func)
 export(DSF_Scale)
 export(DSF_dplyr)

NEWS.md

@@ -1,8 +1,16 @@
 # stream 2.0-2 (xx/xx/23)
 
+## New Features
+* Added DSF_FeatureSelection.
+
 ## Bugfix
 * Fixed exports for get_macroclusters.DSC_DStream, get_macroweights.DSC_DStream
   and recluster.DSC
+* predict for DSC now honors the formula in the DSC.
+
+## Changes
+* Improved manual pages
+* proxy is now imported and not attached globally.
 
 # stream 2.0-1 (02/28/23)
 

R/DSC_DBSTREAM.R

@@ -142,6 +142,19 @@
 #'
 #' update(dbstream, stream, 500)
 #' get_centers(dbstream)
+#'
+#' # use DBSTREAM for outlier detection
+#' stream <- DSD_Gaussians(k = 3, d = 4, noise = 0.05)
+#' outlier_detector <- DSOutlier_DBSTREAM(r = .2)
+#'
+#' update(outlier_detector, stream, 500)
+#' outlier_detector
+#'
+#' plot(outlier_detector, stream)
+#'
+#' points <- get_points(stream, 20)
+#' points
+#' which(is.na(predict(outlier_detector, points)))
 #' @export
 DSC_DBSTREAM <- function(formula = NULL,
   r,
@@ -548,6 +561,10 @@ get_assignment.DSC_DBSTREAM <- function(dsc,
 
   points <- remove_info(points)
 
+  ## apply formula
+  if (!is.null(dsc$RObj$colnames))
+    points <- points[, dsc$RObj$colnames, drop = FALSE]
+
   if (method == "auto")
     method <- "model"
   if (method != "model")

R/DSC_DStream.R

@@ -727,6 +727,10 @@ get_assignment.DSC_DStream <-
 
     points <- remove_info(points)
 
+    ## apply formula
+    if (!is.null(dsc$RObj$colnames))
+      points <- points[, dsc$RObj$colnames, drop = FALSE]
+
     if (method == "auto")
       method <- "model"
     if (method != "model")

R/DSC_R.R

@@ -100,6 +100,7 @@ update.DSC_R <- function(object,
   for (bl in .make_block(n, block)) {
     p <- get_points(dsd, bl, info = FALSE)
 
+    ## apply formula
     if (!is.null(object$formula)) {
       if (is.null(object$RObj$colnames)) {
         trms <- terms(object$formula, data = p)

R/DSC_TwoStage.R

@@ -163,6 +163,10 @@ get_assignment.DSC_TwoStage <-
     ...) {
     points <- remove_info(points)
 
+    ## apply formula
+    if (!is.null(dsc$RObj$colnames))
+      points <- points[, dsc$RObj$colnames, drop = FALSE]
+
     type <- match.arg(type)
     if (type == "micro") {
       dsc$state$newdata <- TRUE

R/get_assignment.R

@@ -93,8 +93,12 @@ get_assignment.DSC <-
       warning("method model not implemented! using Euclidean nearest neighbor instead!")
       method <- "nn"
     }
-
     points <- remove_info(points)
+
+    ## apply formula
+    if (!is.null(dsc$formula) && !is.null(dsc$RObj$colnames))
+        points <- points[, dsc$RObj$colnames, drop = FALSE]
+
     c <- get_centers(dsc, type = type, ...)
 
     if (nrow(c) > 0L) {