Movatterモバイル変換

Title:	Perform Monothetic Clustering with Extensions to Circular Data
Version:	1.2.1
Description:	Implementation of the Monothetic Clustering algorithm (Chavent, 1998 <doi:10.1016/S0167-8655(98)00087-7>) on continuous data sets. A lot of extensions are included in the package, including applying Monothetic clustering on data sets with circular variables, visualizations with the results, and permutation and cross-validation based tests to support the decision on the number of clusters.
License:	GPL-2 |GPL-3 [expanded from: GPL (≥ 2)]
URL:	https://vinhtantran.github.io/monoClust/,https://github.com/vinhtantran/monoClust
BugReports:	https://github.com/vinhtantran/monoClust/issues
Depends:	R (≥ 3.3.0)
Imports:	cluster (≥ 2.0.5), doParallel, dplyr (≥ 1.0.0), foreach,ggplot2, graphics, grDevices, parallel, permute, purrr (≥0.3.0), rlang (≥ 0.3.0), stats, stringr (≥ 0.5), tibble (≥3.0.0), tidyr (≥ 1.0.0)
Suggests:	knitr, mice, rmarkdown, covr, testthat (≥ 3.0.0)
VignetteBuilder:	knitr
Encoding:	UTF-8
LazyData:	true
RoxygenNote:	7.1.1
Config/testthat/edition:	3
NeedsCompilation:	no
Packaged:	2021-02-15 00:10:00 UTC; vinht
Author:	Tan Tran [aut, cre], Brian McGuire [aut], Mark Greenwood [aut]
Maintainer:	Tan Tran <vinhtantran@gmail.com>
Repository:	CRAN
Date/Publication:	2021-02-15 15:00:02 UTC

monoClust: Perform Monothetic Clustering with Extensions to CircularData

Description

Implementation of the Monothetic Clusteringalgorithm (Chavent, 1998 <doi:10.1016/S0167-8655(98)00087-7>) oncontinuous data sets. A lot of extensions are included in the package,including applying Monothetic clustering on data sets with circularvariables, visualizations with the results, and permutation andcross-validation based tests to support the decision on the number ofclusters.

Author(s)

Maintainer: Tan Tranvinhtantran@gmail.com (ORCID)

Authors:

• Brian McGuiremcguirebc@gmail.com

• Mark Greenwoodgreenwood@montana.edu (ORCID)

See Also

Useful links:

• https://vinhtantran.github.io/monoClust/

• https://github.com/vinhtantran/monoClust

• Report bugs athttps://github.com/vinhtantran/monoClust/issues

Monothetic Clustering

Description

Creates a MonoClust object after partitioning the data set using MonotheticClustering.

Usage

MonoClust(  toclust,  cir.var = NULL,  variables = NULL,  distmethod = NULL,  digits = getOption("digits"),  nclusters = 2L,  minsplit = 5L,  minbucket = round(minsplit/3),  ncores = 1L)

Arguments

Value

AMonoClust object. SeeMonoClust.object.

References

1. Chavent, M. (1998). A monothetic clustering method. Pattern RecognitionLetters, 19(11), 989-996. doi:10.1016/S0167-8655(98)00087-7.

2. Tran, T. V. (2019). Monothetic Cluster Analysis with Extensions toCircular and Functional Data. Montana State University - Bozeman.

Examples

# Very simple data setlibrary(cluster)data(ruspini)ruspini4sol <- MonoClust(ruspini, nclusters = 4)ruspini4sol# data with circular variablelibrary(monoClust)data(wind_sensit_2007)# Use a small data setset.seed(12345)wind_reduced <- wind_sensit_2007[sample.int(nrow(wind_sensit_2007), 10), ]circular_wind <- MonoClust(wind_reduced, cir.var = 3, nclusters = 2)circular_wind

Monothetic Clustering Tree Object

Description

The structure and objects contained in MonoClust, an object returned fromtheMonoClust() function and used as the input in other functions in thepackage.

Value

frame

Data frame in the form of atibble::tibble() representinga tree structure with one row for each node. The columns include:

number

Index of the node. Depth of a node can be derived bynumber %/% 2.

var

Name of the variable used in the split at a node or"<leaf>" if it is a leaf node.

cut

Splitting value, so values ofvar that are smaller thanthat go to left branch while values greater than that go to the rightbranch.

n

Cluster size, the number of observations in that cluster.

inertia

Inertia value of the cluster at that node.

bipartsplitrow

Position of the next split row in the data set(that position will belong to left node (smaller)).

bipartsplitcol

Position of the next split variable in the dataset.

inertiadel

Proportion of inertia value of the cluster at thatnode to the inertia of the root.

medoid

Position of the data point regarded as the medoid ofits cluster.

loc

y-coordinate of the splitting node to facilitate showingon the tree. Seeplot.MonoClust() for details.

split.order

Order of the splits with root is 0.

inertia_explained

Percent inertia explained as described inChavent (2007). It is⁠1 - (sum(current inertia)/inertial[1])⁠.

alt

A nested tibble of alternate splits at a node. It containsbipartsplitrow andbipartsplitcol with the same meaning above.Note that this is only for information purpose. CurrentlymonoClustdoes not support choosing an alternate splitting route. RunningMonoClust() withnclusters = 2 step-by-step can be run ifneeded.

membership

Vector of the same length as the number of rows in thedata, containing the value offrame$number corresponding to the leafnode that an observation falls into.

dist

Distance matrix calculated using the method indicated indistmethod argument ofMonoClust().

terms

Vector of variable names in the data that were used to split.

centroids

Data frame with one row for centroid value of eachcluster.

medoids

Named vector of positions of the data points regarded asmedoids of clusters.

alt

Indicator of having an alternate splitting route occurred whensplitting.

circularroot

List of values designed for circular variable in thedata set.var is the name of circular variable andcut is its firstbest split value. If circular variable is not available, both objects areNULL.

References

• Chavent, M., Lechevallier, Y., & Briant, O. (2007). DIVCLUS-T: A monotheticdivisive hierarchical clustering method. Computational Statistics & DataAnalysis, 52(2), 687-701. doi:10.1016/j.csda.2007.03.013.

See Also

MonoClust().

Coerce Similar Object to MonoClust

Description

The function turns a MonoClust-similar object into MonoClust object so itcan use supported functions for MonoClust such asprint.MonoClust() andplot.MonoClust().

Usage

as_MonoClust(x, ...)## Default S3 method:as_MonoClust(x, ...)

Arguments

Details

as_MonoClust() is an S3 generic. The function itself doesn't run unlessit is implemented for another similar object. Currently, this function is notimplemented withinmonoClust package.

Find Centroid of the Cluster

Description

Centroid is point whose coordinates are the means of their cluster values.

Usage

centroid(data, frame, cloc)

Arguments

Value

A data frame with coordinates of centroids

First Gate Function

Description

This function checks what are available nodes to split and then callfind_split() on each node, then decide which node creates best split, andcallsplitter() to perform the split.

Usage

checkem(  data,  cuts,  frame,  cloc,  dist,  variables,  minsplit,  minbucket,  split_order,  ncores)

Arguments

Value

It is not supposed to return anything because global environment wasused. However, if there is nothing left to split, it returns 0 to tell thecaller to stop running the loop.

Add/Subtract Circular Values in Degrees/Radian

Description

Add/subtract two circular variables in degrees (⁠%cd+%⁠ and⁠%cd-%⁠) andradian (⁠%cr+%⁠ and⁠%cr-%⁠).

Usage

x %cd+% yx %cd-% yx %cr+% yx %cr-% y

Arguments

Value

A value between [0, 360) in degrees or [0, 2*pi) in radian.

Examples

90 %cd+% 90250 %cd+% 20025 %cd-% 80pi %cr+% (pi/2)

Distance Matrix of Circular Variables

Description

Calculates the distance matrix of observations with circular variables usingan adapted version of Gower's distance. This distance should be compatiblewith the Gower's distance for other variable types.

Usage

circ_dist(frame)

Arguments

Details

The distance between two observations i and j of a circular variable qis suggested to be

(y_{iq}, y_{jq}) = \frac{180 - |180 - |y_{iq} - y_{jq}||}{180}.

Value

Object of class "dist".

References

• Tran, T. V. (2019). Chapter 3. Monothetic Cluster Analysis with Extensionsto Circular and Functional Data. Montana State University - Bozeman.

See Also

stats::dist()

Examples

# Make a sample data set of 20 observations with 2 circular variablesdata <- data.frame(var1 = sample.int(359, 20),                   var2 = sample.int(359, 20))circ_dist(data)

Cluster Statistics Calculation

Description

Calinski-Harabasz's pseudo-F (Calinski and Harabasz, 1974) and Averagesilhoutte width (Rousseeuw, 1987) calculation.

Usage

cluster_stats(d, clustering)

Arguments

Value

f_stat

Calinski-Harabasz's pseudo-F.

asw

Average silhouette width.

References

• Caliński, T. and Harabasz, J (1974). "A dendrite method for clusteranalysis". en. In:Communications in Statistics 3.1, pp. 1–27.doi:10.1080/03610927408827101.

• Rousseeuw, P. J. (1987). "Silhouettes: A graphical aid to theinterpretation and validation of cluster analysis". In:Journal ofComputational and Applied Mathematics 20, pp. 53–65. ISSN: 03770427.doi:10.1016/0377-0427(87)90125-7.

See Also

cluster::silhouette()

Create Labels for Split Variables

Description

This function prints variable's labels for aMonoClust tree.

Usage

create_labels(x, abbrev, digits = getOption("digits"), ...)

Arguments

Value

A list containing two elements:

• varnames: A named vector of labels corresponding to variable's names(at vector names).

• labels: Vector of labels of splitting rules to be displayed.

See Also

abbreviate()

Cross-Validation Test on MonoClust

Description

Perform cross-validation test for different different number of clusters ofMonothetic Clustering.

Usage

cv.test(data, fold = 10L, minnodes = 2L, maxnodes = 10L, ncores = 1L, ...)

Arguments

Details

Thek-fold cross-validation randomly partitions data intoksubsets with equal (or close to equal) sizes.k - 1 subsets are used asthe training data set to create a tree with a desired number of leaves andthe other subset is used as validation data set to evaluate the predictiveperformance of the trained tree. The process repeats for each subset as thevalidating set (m = 1, \ldots, k) and the mean squared difference,

MSE_m=\frac{1}{n_m} \sum_{q=1}^Q\sum_{i \in m} d^2_{euc}(y_{iq},\hat{y}_{(-i)q}),

is calculated, where\hat{y}_{(-i)q} is the cluster mean on thevariableq of the cluster created by the training data where the observed value,y_{iq}, of the validation data set will fall into, andd^2_{euc}(y_{iq}, \hat{y}_{(-i)q}) is the squared Euclidean distance(dissimilarity) between two observations at variable $q$. This process isrepeated for the $k$ subsets of the data set and the average of these testerrors is the cross-validation-based estimate of the mean squared error ofpredicting a new observation,

CV_K = \overline{MSE} = \frac{1}{M} \sum_{m=1}^M MSE_m.

Value

AMonoClust.cv class containing a data frame of mean sum of squareerror and its standard deviation.

Note

This function supports parallel processing withforeach::foreach().It distributes MonoClust calls to processes.

See Also

plot.cv.MonoClust(),MonoClust(),predict.MonoClust()

Examples

library(cluster)data(ruspini)# Leave-one-out cross-validationcv.test(ruspini, fold = 1, minnodes = 2, maxnodes = 4)# 5-fold cross-validationcv.test(ruspini, fold = 5, minnodes = 2, maxnodes = 4)

Make Error Bars

Description

Make Error Bars

Usage

error_bar(x, y, upper, lower = upper, length = 0.1, ...)

Arguments

Value

Plot

Find the Closest Cut

Description

Find the cuts for a quantitative variable. These cuts are what we aregoing to consider when thinking about bi-partitioning the data. For aquantitative column, find the next larger value of each value, if it is thelargest, that value + 1

Usage

find_closest(col)

Arguments

Value

a quantitative vector which contains the closest higher cut.

Find the Best Split

Description

Find the best split in terms of reduction in inertia for the transferrednode, indicate by row. Find the terminal node with the greatest change ininertia and bi-partition it.

Usage

find_split(  data,  cuts,  frame_row,  cloc,  dist,  variables,  minsplit,  minbucket,  ncores)

Arguments

Value

The updatedframe_row with the next split updated.

GGPlot the Mean Square Error with Error Bar for +/- 1 Standard Error

Description

GGPlot the Mean Square Error with Error Bar for +/- 1 Standard Error

Usage

ggcv(  cv.obj,  title = "MSE for CV of monothetic clustering",  xlab = "Number of clusters",  ylab = "MSE +/- 1 SE",  type = c("b", "p", "l"),  linetype = 2,  err.col = "red",  err.width = 0.2)

Arguments

Value

A ggplot2 object.

See Also

Plot using base Rplot.cv.MonoClust()

Examples

library(cluster)data(ruspini)# 10-fold cross-validationcptable <- cv.test(ruspini, minnodes = 2, maxnodes = 4)ggcv(cptable)

Parallel Coordinates Plot with Circular Variables

Description

Making a parallel coordinates plot with the circular variables are plottedas ellipses. The function currently works well with data with one circularvariable.

Usage

ggpcp(  data,  circ.var = NULL,  is.degree = TRUE,  rotate = 0,  north = 0,  cw = FALSE,  order.appear = NULL,  linetype = 1,  size = 0.5,  alpha = 0.5,  clustering,  medoids = NULL,  cluster.col = NULL,  show.medoids = FALSE,  labelsize = 4,  xlab = "Variables",  ylab = NULL,  legend.cluster = "groups")

Arguments

Value

A ggplot2 object.

Examples

# Set color constantCOLOR4 <- c("#e41a1c", "#377eb8", "#4daf4a", "#984ea3")# Reduce the size of the data for for sake of example speedset.seed(12345)wind_reduced <- wind_sensit_2007[sample.int(nrow(wind_sensit_2007), 50), ]sol42007 <- MonoClust(wind_reduced, cir.var = 3, nclusters = 4)library(ggplot2)ggpcp(data = wind_reduced,      circ.var = "WDIR",      # To improve aesthetics      rotate = pi*3/4-0.3,      order.appear = c("WDIR", "has.sensit", "WS"),      alpha = 0.5,      clustering = sol42007$membership,      medoids = sol42007$medoids,      cluster.col = COLOR4,      show.medoids = TRUE) +  theme(panel.background = element_rect(color = "white"),        panel.border = element_rect(color = "white", fill = NA),        panel.grid.major = element_line(color = "#f0f0f0"),        panel.grid.minor = element_blank(),        axis.line = element_line(color = "black"),        legend.key = element_rect(color = NA),        legend.position = "bottom",        legend.direction = "horizontal",        legend.title = element_text(face = "italic"),        legend.justification = "center")

Cluster Inertia Calculation

Description

Calculate inertia for a given subset of the distance matrix from the originaldata set provided tox. Assumes that distance matrices are stored asmatrices and not distance objects.

Usage

inertia_calc(x)

Arguments

Value

Inertia value of the matrix, formula in Chavent (1998). Ifx is asingle number, return 0.

Examples

data(iris)# Euclidean distance on first 20 rows of the 4 continuous variablesdist_mat <- as.matrix(dist(iris[1:20, 1:4]))inertia_calc(dist_mat)

Test If The Object is A MonoClust

Description

This function returnsTRUE for MonoClust, and FALSE for all other objects.

Usage

is_MonoClust(mono_obj)

Arguments

Value

TRUE if the object inherits from theMonoClust class.

Create Jump Table

Description

Create jump table from the MonoClust's frame object.number andvar willbe used to create the table.

Usage

make_jump_table(frame)

Arguments

Value

Jump table withnumber,var, and two new columnsleft andright indicate the left and right number at split.

Find Medoid of the Cluster

Description

Medoid is the point that has minimum distance to all other points in thecluster.

Usage

medoid(members, dist_mat)

Arguments

Value

index of the medoid point in the members vector.

Examples

library(cluster)data(ruspini)ruspini4sol <- MonoClust(ruspini, nclusters = 4)ruspini4solmedoid(which(ruspini4sol$membership == 4), ruspini4sol$dist)# Check with the output with "4" labelruspini4sol$medoids

Create A New Node for Split Data Frame

Description

This function is just a helper to make sure that the default values of thesplit data frame is correct when unspecified. It helps reduce type error,especially when moving to use dplyr which is stricter in data types.

Usage

new_node(  number,  var,  cut = -99L,  n,  inertia,  bipartsplitrow = -99L,  bipartsplitcol = -99L,  inertiadel = 0,  inertia_explained = -99,  medoid,  loc,  split.order = -99L,  alt = list(tibble::tibble(bipartsplitrow = numeric(), bipartsplitcol = numeric())))

Arguments

Value

A tibble with only one row and correct default data type for even anunspecified variables.

References

• Chavent, M., Lechevallier, Y., & Briant, O. (2007). DIVCLUS-T: A monotheticdivisive hierarchical clustering method. Computational Statistics & DataAnalysis, 52(2), 687–701. https://doi.org/10.1016/j.csda.2007.03.013

Permutation Test on Monothetic Tree

Description

Testing the significance of each monothetic clustering split by permutationmethods. The "simple-withhold" method ("sw") shuffles the observationsbetween two groups without the splitting variable. The other two methodsshuffle the values in the splitting variable to create a new data set, thenit either splits again on that variable ("resplit-limit","rl") or use allvariables as the splitting candidates ("resplit-nolimit","rn").

Usage

perm.test(  object,  data,  auto.pick = FALSE,  sig.val = 0.05,  method = c("sw", "rl", "rn"),  rep = 1000L,  stat = c("f", "aw"),  bon.adj = TRUE,  ncores = 1L)

Arguments

Details

Permutation Methods

Simple-Withhold: Shuffle the observations between two proposed clusters

Thestat calculated from the shuffles create the reference distributionto find the p-value. Because the splitting variable that was chosen isalready the best in terms of reduction of inertia, that variable is withheldfrom the distance matrix used in the permutation test.

Resplit-Limit: Shuffle splitting variable, split again on that variable

This method shuffles the values of the splitting variables while keepingother variables fixed to create a new data set, then the chosenstat iscalculated for each rep to compare with the observedstat.

Resplit-Nolimit: Shuffle splitting variable, split on all variables

Similar to Method 2 but all variables are splitting candidates.

Bonferroni Correction

A hypothesis test occurred lower in the monothetic clustering tree could haveits p-value corrected for multiple tests happened before it in order to reachthat node. The formula is

adj.p = unadj.p \times depth,

withdepth is 1 at the root node.

Value

The sameMonoClust object with an extra column (p-value), as wellas thenumofclusters object ifauto.pick = TRUE.

Note

This function usesforeach::foreach() to facilitate parallelprocessing. It distributes reps to processes.

References

Calinski, T. and Harabasz, J (1974). "A dendrite method for clusteranalysis". en. In:Communications in Statistics 3.1, pp. 1-27.doi:10.1080/03610927408827101.

Rousseeuw, P. J. (1987). "Silhouettes: A graphical aid to the interpretationand validation of cluster analysis". In:Journal of Computational andApplied Mathematics 20, pp. 53-65. ISSN: 03770427.doi:10.1016/0377-0427(87)90125-7.

Examples

library(cluster)data(ruspini)ruspini6sol <- MonoClust(ruspini, nclusters = 6)ruspini6.p_value <- perm.test(ruspini6sol, data = ruspini, method = "sw",                              rep = 1000)ruspini6.p_value

Plot MonoClust Splitting Rule Tree

Description

Print the MonoClust tree in the form of dendrogram.

Usage

## S3 method for class 'MonoClust'plot(  x,  uniform = FALSE,  branch = 1,  margin = c(0.12, 0.02, 0, 0.05),  minbranch = 0.3,  text = TRUE,  which = 4,  stats = TRUE,  abbrev = c("no", "short", "abbreviate"),  digits = getOption("digits") - 2,  cols = NULL,  col.type = c("l", "p", "b"),  rel.loc.x = TRUE,  show.pval = TRUE,  ...)

Arguments

Value

A plot of splitting rule.

Examples

library(cluster)data(ruspini)# MonoClust treeruspini4sol <- MonoClust(ruspini, nclusters = 4)plot(ruspini4sol)# MonoClust tree after permutation test is runruspini6sol <- MonoClust(ruspini, nclusters = 6)ruspini6_test <- perm.test(ruspini6sol,                           data = ruspini,                           method = "sw",                           rep = 1000)plot(ruspini6_test, branch = 1, uniform = TRUE)

Plot the Mean Square Error with Error Bar for +/- 1 Standard Error

Description

Plot the Mean Square Error with Error Bar for +/- 1 Standard Error

Usage

## S3 method for class 'cv.MonoClust'plot(  x,  main = "MSE for CV of monothetic clustering",  xlab = "Number of clusters",  ylab = "MSE +/- 1 SE",  type = "b",  lty = 2,  err.col = "red",  err.width = 0.1,  ...)

Arguments

Value

A line plot with error bars.

See Also

Plot using ggplot2ggcv()

Examples

library(cluster)data(ruspini)# 10-fold cross-validationcptable <- cv.test(ruspini, minnodes = 2, maxnodes = 4)plot(cptable)

Calculate Branch Coordinates

Description

Calculate Branch Coordinates

Usage

plot_prep_branch(x, y, node, branch = 0)

Arguments

Value

Branch coordinates in a list of x and y axis.

Calculate Nodes Coordinates

Description

Calculate Nodes Coordinates

Usage

plot_prep_node(tree, uniform = FALSE, minbranch = 0.3)

Arguments

Value

Nodes coordinates in a list of x and y axis.

Plot the monoClust Tree.

Description

This function plots the MonoClust tree. It is partially inspired by rpartpackage.

Usage

plot_tree(  x,  uniform = FALSE,  branch = 1,  margin = 0,  minbranch = 0.3,  rel.loc.x = TRUE,  ...)

Arguments

Value

Plot of tree

Predictions from a MonoClust Object

Description

Predict the cluster memberships of a new data set from aMonoClust object.

Usage

## S3 method for class 'MonoClust'predict(object, newdata, type = c("centroid", "medoid"), ...)

Arguments

Value

A tibble of cluster index incname and either centroid values ormedoid observations index based on the value oftype argument.

Examples

library(cluster)data(ruspini)set.seed(1234)test_index <- sample(1:nrow(ruspini), nrow(ruspini)/5)train_index <- setdiff(1:nrow(ruspini), test_index)ruspini_train <- ruspini[train_index, ]ruspini_test <- ruspini[test_index, ]ruspini_train_4sol <- MonoClust(ruspini_train, nclusters = 4)predict(ruspini_train_4sol, newdata = ruspini_test)

Print Monothetic Clustering Results

Description

Render theMonoClust split tree in an easy to read format with importantinformation such as terminal nodes, p-value (if possible), etc.

Usage

## S3 method for class 'MonoClust'print(  x,  abbrev = c("no", "short", "abbreviate"),  spaces = 2L,  digits = getOption("digits"),  ...)

Arguments

Value

A nicely displayed MonoClust split tree.

See Also

abbreviate()

Examples

library(cluster)data(ruspini)ruspini4sol <- MonoClust(ruspini, nclusters = 4)print(ruspini4sol, digits = 2)

Print MonoClust Cross-Validation Result

Description

Print MonoClust Cross-Validation Result

Usage

## S3 method for class 'cv.MonoClust'print(x, ...)

Arguments

Examples

library(cluster)data(ruspini)# 10-fold cross-validationcp_table <- cv.test(ruspini, minnodes = 2, maxnodes = 4)print(cp_table)

Split Function

Description

Given the Cluster's frame's row position to split atsplit_row, thisfunction performs the split, calculate all necessary information for thesplitting tree and cluster memberships.

Usage

splitter(data, cuts, split_row, frame, cloc, dist, split_order = 0L)

Arguments

Value

Updatedframe andcloc saved in a list.

Hypothesis Test at Split

Description

Hypothesis Test at Split

Usage

test_split(members_l, members_r, method, data, split_var, rep, stat, ncores)

Arguments

Value

p-value of the test

Implementation of Print Labels on MonoClust Tree

Description

This function plots the labels onto the MonoClust tree. It is partiallyinspired by rpart package.

Usage

text_tree(  x,  which = 4,  digits = getOption("digits") - 2,  stats = TRUE,  abbrev,  cols = NULL,  cols.type = c("l", "p", "b"),  rel.loc.x = TRUE,  show.pval = TRUE,  uniform = FALSE,  minbranch = 0.3,  ...)

Arguments

Value

Labels on tree.

Transform Between Degree and Radian

Description

This function transforms a circular angle from degree to radian or fromradian to degree.

Usage

torad(x)todeg(x)

Arguments

Value

A radian value iftorad or degree value iftodeg.

Examples

torad(90)torad(-45)todeg(pi/2)

Find Tree Depth Based on Node Indexes

Description

Find Tree Depth Based on Node Indexes

Usage

tree_depth(nodes)

Arguments

Details

When building MonoClust tree, the node index was created with the rule thatnew node indexes are the split node times 2 plus 0 (left) and 1 (right).Therefore, this function is just a back-transform, taking a log base 2.

Value

Depth of the node, with 0 is the root relative to the input.

Traverse a Tree to Find the Leaves (Terminal Nodes)

Description

Traverse a Tree to Find the Leaves (Terminal Nodes)

Usage

tree_walk(new_point, jump_table)

Arguments

Value

The index of the terminal node after traversing the new data point onthe tree.

Existence of Microorganisms Carried in Wind

Description

Data set is a part of a study on microorganisms carried in strong f\"ohnwinds at the Bonney Riegel location of Taylor Valley, an ice free area in theAntarctic continent. Wind direction and wind speed data were obtained fromthe meteorological station. Wind direction was recorded every 30 seconds andwind speeds every 4 seconds at 1.15 meters above the ground surface. Therecorded wind directions and speeds were averaged at 15 minute intervals. Forwind direction, as discussed previously, winds from the north are defined as0/360 degrees and from the east as 90 degrees. 2007 data were collected fromAugust 4–11, 2007.

Usage

wind_sensit_2007

Format

A data frame with 671 rows and 3 variables:

has.sensit

A binary variable of the existence of particles in thewind (1) or not (0).

WS

Wind speed measured in m/s.

WDIR

Wind direction in degree with 0 indicates "from the north" and90 degrees indicate "from the east".

Source

Sabacka, M., Priscu, J. C., Basagic, H. J., Fountain, A. G., Wall, D.H., Virginia, R. A., and Greenwood, M. C. (2012). "Aeolian flux of biotic andabiotic material in Taylor Valley, Antarctica". In: Geomorphology 155-156,pp. 102-111. issn: 0169555X. doi:10.1016/j.geomorph.2011.12.009.

Existence of Microorganisms Carried in Wind

Description

Data set is a part of a study on microorganisms carried in strong f\"ohnwinds at the Bonney Riegel location of Taylor Valley, an ice free area in theAntarctic continent. Wind direction and wind speed data were obtained fromthe meteorological station. Wind direction was recorded every 30 seconds andwind speeds every 4 seconds at 1.15 meters above the ground surface. Therecorded wind directions and speeds were averaged at 15 minute intervals. Forwind direction, as discussed previously, winds from the north are defined as0/360 degrees and from the east as 90 degrees. 2008 data were collected fromJuly 7–14, 2008.

Usage

wind_sensit_2008

Format

A data frame with 673 rows and 3 variables:

has.sensit

A binary variable of the existence of particles in thewind (1) or not (0).

WS

Wind speed measured in m/s.

WDIR

Wind direction in degree with 0 indicates "from the north" and90 degrees indicate "from the east".

Source

Sabacka, M., Priscu, J. C., Basagic, H. J., Fountain, A. G., Wall, D.H., Virginia, R. A., and Greenwood, M. C. (2012). "Aeolian flux of biotic andabiotic material in Taylor Valley, Antarctica". In: Geomorphology 155-156,pp. 102-111. issn: 0169555X. doi:10.1016/j.geomorph.2011.12.009.