### Install and Load ClustOfVar

Source: https://context7.com/chavent/clustofvar/llms.txt

Install the package from GitHub and load it into the R environment.

```r
# Install from GitHub
devtools::install_github("chavent/ClustOfVar")

# Load the package
library(ClustOfVar)
```

--------------------------------

### Install ClustOfVar from GitHub

Source: https://github.com/chavent/clustofvar/blob/master/README.md

Use devtools to install the development version of the package from the GitHub repository.

```r
devtools::install_github("chavent/ClustOfVar")
# This needs the devtools package to be installed :
# install.packages("devtools")
```

--------------------------------

### Evaluate Partition Stability

Source: https://context7.com/chavent/clustofvar/llms.txt

Use bootstrap resampling to assess the stability of hierarchical clustering partitions.

```r
library(ClustOfVar)
data(decathlon)

tree <- hclustvar(X.quanti = decathlon[, 1:10])

stab <- stability(tree, B = 60, graph = TRUE)

stab$matCR    # matrix of corrected Rand indices (B x (p-2))
stab$meanCR   # vector of mean corrected Rand indices

plot(stab, nmax = 7)

boxplot(stab$matCR[, 1:7])

optimal_k <- which.max(stab$meanCR) + 1
cat("Suggested number of clusters:", optimal_k)
```

--------------------------------

### Plot Variable Loadings by Cluster

Source: https://context7.com/chavent/clustofvar/llms.txt

Visualize variable contributions to clusters using dot charts.

```r
library(ClustOfVar)
data(wine)

X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali
tree <- hclustvar(X.quanti, X.quali)
tree_cut <- cutreevar(tree, k = 6)

res_plot <- plot(tree_cut)

res_plot$coord.quanti  # coordinates of quantitative variables
res_plot$coord.levels  # coordinates of categorical variable levels
```

--------------------------------

### Perform K-Means Clustering of Variables

Source: https://context7.com/chavent/clustofvar/llms.txt

Use kmeansvar to perform iterative k-means clustering, supporting random initialization or initialization from a hierarchical partition.

```r
library(ClustOfVar)
data(decathlon)

# K-means with random initialization, 10 random starts
part_random <- kmeansvar(
  X.quanti = decathlon[, 1:10],
  init = 5,           # number of clusters
  nstart = 10,        # number of random initializations
  iter.max = 150      # maximum iterations
)
summary(part_random)

# Output:
# Call:
# kmeansvar(X.quanti = decathlon[, 1:10], init = 5, nstart = 10)
#
# number of iterations: 3
#
# Data:
#    number of observations:  41
#    number of variables:  10
#    number of clusters:  5
#
# Cluster 1:
#          squared loading correlation
# X100m              0.92       -0.96
# Long.jump          0.88        0.94
# ...

# K-means initialized from hierarchical clustering
tree <- hclustvar(X.quanti = decathlon[, 1:10])
part_init <- cutreevar(tree, 5)$cluster
part_hier <- kmeansvar(
  X.quanti = decathlon[, 1:10],
  init = part_init,   # use partition as initial clusters
  matsim = TRUE       # calculate similarity matrices
)
summary(part_hier)
part_hier$sim  # view similarity matrices

# Access cluster assignments
part_random$cluster
# X100m  Long.jump   Shot.put  High.jump      X400m
#     1          1          2          3          4
# ...

# Access cluster scores (synthetic variables)
head(part_random$scores)
#           cluster1   cluster2   cluster3   cluster4   cluster5
# SEBRLE   0.4872419 -0.2551023 -0.5813429 -0.8629382  0.5019456
# ...
```

--------------------------------

### Plot Dendrogram and Aggregation Levels

Source: https://context7.com/chavent/clustofvar/llms.txt

Visualize hierarchical clustering results as a dendrogram or aggregation index plot.

```r
library(ClustOfVar)
data(wine)

X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali
tree <- hclustvar(X.quanti, X.quali)

plot(tree)

plot(tree, type = "index")
```

--------------------------------

### Access ClustOfVar Partition Results

Source: https://context7.com/chavent/clustofvar/llms.txt

Access cluster assignments, sizes, and homogeneity metrics from a clustering object.

```r
summary(part)

part$cluster     # cluster assignments
part$size        # number of variables per cluster
part$wss         # within-cluster sum of squares
part$E           # percentage of homogeneity explained

part_sim <- cutreevar(tree, k = 4, matsim = TRUE)
part_sim$sim$cluster1
```

--------------------------------

### Cut Hierarchical Tree into Clusters

Source: https://context7.com/chavent/clustofvar/llms.txt

Use cutreevar to partition a hierarchical tree into a specific number of clusters and retrieve cluster details.

```r
library(ClustOfVar)
data(decathlon)

# Build hierarchical tree
tree <- hclustvar(decathlon[, 1:10])
plot(tree)

# Determine optimal number of clusters using stability
stab <- stability(tree, B = 40)

# Cut tree into 4 clusters
part <- cutreevar(tree, k = 4)
print(part)

# Output:
# Call:
# cutreevar(obj = tree, k = 4)
#
#         name      description
#  "$var"    "list of variables in each cluster"
#  "$sim"    "similarity matrix in each cluster"
#  "$cluster""cluster memberships"
#  "$wss"    "within-cluster sum of squares"
#  "$E"      "gain in cohesion (in %)"
#  "$size"   "size of each cluster"
#  "$scores" "synthetic score of each cluster"
```

--------------------------------

### plot.clustvar - Plot Variable Loadings by Cluster

Source: https://context7.com/chavent/clustofvar/llms.txt

Creates dot charts showing the loadings of variables within each cluster. For quantitative variables, loadings are correlations with the synthetic variable; for categorical variables, loadings are mean values of the synthetic variable for each level.

```APIDOC
## plot.clustvar - Plot Variable Loadings by Cluster

### Description
Creates dot charts showing the loadings of variables within each cluster. For quantitative variables, loadings are correlations with the synthetic variable; for categorical variables, loadings are mean values of the synthetic variable for each level.

### Method
`plot(x, ...)`

### Parameters
#### Path Parameters
None

#### Query Parameters
- **x** (clustvar object) - The result of a clustering function (e.g., `cutreevar`).

### Request Example
```r
library(ClustOfVar)
data(wine)

# Prepare and cluster mixed data
X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali
tree <- hclustvar(X.quanti, X.quali)
tree_cut <- cutreevar(tree, k = 6)

# Plot loadings for each cluster
res_plot <- plot(tree_cut)
```

### Response
#### Success Response (200)
- Generates a plot showing variable loadings by cluster.
- **coord.quanti** (matrix) - Coordinates of quantitative variables.
- **coord.levels** (matrix) - Coordinates of categorical variable levels.

#### Response Example
```r
# Plots are displayed directly.
# Access coordinate data
# res_plot$coord.quanti
# res_plot$coord.levels
```
```

--------------------------------

### stability - Evaluate Partition Stability

Source: https://context7.com/chavent/clustofvar/llms.txt

Evaluates the stability of partitions obtained from hierarchical clustering using a bootstrap approach. The function compares partitions from bootstrap samples with the original partition using the adjusted Rand index, helping to determine an appropriate number of clusters.

```APIDOC
## stability - Evaluate Partition Stability

### Description
Evaluates the stability of partitions obtained from hierarchical clustering using a bootstrap approach. The function compares partitions from bootstrap samples with the original partition using the adjusted Rand index, helping to determine an appropriate number of clusters.

### Method
`stability(tree, B, graph, adj, method, ...)`

### Parameters
#### Path Parameters
None

#### Query Parameters
- **tree** (hclustvar object) - The hierarchical clustering result.
- **B** (numeric) - The number of bootstrap samples.
- **graph** (logical) - If TRUE, a plot is generated.
- **adj** (logical) - If TRUE, uses adjusted Rand index.
- **method** (character) - Method for calculating similarity (e.g., "average", "complete").

### Request Example
```r
library(ClustOfVar)
data(decathlon)

# Build hierarchical tree
tree <- hclustvar(X.quanti = decathlon[, 1:10])

# Evaluate stability with 60 bootstrap samples
stab <- stability(tree, B = 60, graph = TRUE)
```

### Response
#### Success Response (200)
- **matCR** (matrix) - Matrix of corrected Rand indices (B x (p-2)).
- **meanCR** (vector) - Vector of mean corrected Rand indices.

#### Response Example
```r
# Access stability results
stab$matCR
stab$meanCR

# Custom plot with limited clusters
plot(stab, nmax = 7)

# Boxplot of Rand indices for first 7 partitions
boxplot(stab$matCR[, 1:7])

# Use stability to choose optimal k
optimal_k <- which.max(stab$meanCR) + 1
cat("Suggested number of clusters:", optimal_k)
```
```

--------------------------------

### plot.hclustvar - Plot Dendrogram and Aggregation Levels

Source: https://context7.com/chavent/clustofvar/llms.txt

Visualizes hierarchical clustering results as a dendrogram or as an aggregation levels plot showing the criterion values at each merge step.

```APIDOC
## plot.hclustvar - Plot Dendrogram and Aggregation Levels

### Description
Visualizes hierarchical clustering results as a dendrogram or as an aggregation levels plot showing the criterion values at each merge step.

### Method
`plot(x, type, ...)`

### Parameters
#### Path Parameters
None

#### Query Parameters
- **x** (hclustvar object) - The hierarchical clustering result.
- **type** (character) - Type of plot to draw. "dendrogram" for standard dendrogram, "index" for aggregation levels.

### Request Example
```r
library(ClustOfVar)
data(wine)

# Prepare mixed data
X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali
tree <- hclustvar(X.quanti, X.quali)

# Standard dendrogram
plot(tree)

# Aggregation levels plot
plot(tree, type = "index")
```

### Response
#### Success Response (200)
- Generates a plot.

#### Response Example
```r
# Plots are displayed directly.
```
```

--------------------------------

### Perform Hierarchical Clustering of Variables

Source: https://context7.com/chavent/clustofvar/llms.txt

Use hclustvar to perform hierarchical clustering on quantitative, qualitative, or mixed variable sets.

```r
library(ClustOfVar)

# Clustering quantitative variables only
data(decathlon)
tree_quant <- hclustvar(X.quanti = decathlon[, 1:10])
plot(tree_quant)

# Output structure:
# tree_quant$height   - aggregation criterion values
# tree_quant$merge    - merge matrix showing clustering steps
# tree_quant$clusmat  - cluster membership matrix for all levels

# Clustering qualitative variables with missing values
data(vnf)
tree_qual <- hclustvar(X.quali = vnf)
plot(tree_qual)

# Clustering mixed variables (quantitative + qualitative)
data(wine)
X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali
tree_mixed <- hclustvar(X.quanti, X.quali)
plot(tree_mixed)

# View aggregation levels to help choose number of clusters
plot(tree_mixed, type = "index")
```

--------------------------------

### Calculate Similarity Between Variables

Source: https://context7.com/chavent/clustofvar/llms.txt

Compute squared cosine similarity measures for quantitative, qualitative, or mixed variable pairs.

```r
library(ClustOfVar)
data(wine)

X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali

sim_quant <- mixedVarSim(X.quanti[, 1], X.quanti[, 2])
cat("Quantitative-Quantitative similarity:", sim_quant, "\n")

sim_qual <- mixedVarSim(X.quali[, 1], X.quali[, 2])
cat("Qualitative-Qualitative similarity:", sim_qual, "\n")

sim_mixed <- mixedVarSim(X.quanti[, 1], X.quali[, 1])
cat("Mixed similarity:", sim_mixed, "\n")
```

--------------------------------

### predict.clustvar - Predict Scores for New Data

Source: https://context7.com/chavent/clustofvar/llms.txt

Computes scores for new observations on the synthetic variables (cluster centers) of an existing partition. Uses the linear combination coefficients from the original clustering to project new data onto cluster-specific synthetic variables.

```APIDOC
## predict.clustvar - Predict Scores for New Data

### Description
Computes scores for new observations on the synthetic variables (cluster centers) of an existing partition. Uses the linear combination coefficients from the original clustering to project new data onto cluster-specific synthetic variables.

### Method
`predict(object, X.quanti, X.quali, ...)`

### Parameters
#### Path Parameters
None

#### Query Parameters
- **object** (clustvar object) - The result of a clustering function (e.g., `kmeansvar`).
- **X.quanti** (data.frame) - Quantitative variables in the new data.
- **X.quali** (data.frame) - Qualitative variables in the new data.

### Request Example
```r
library(ClustOfVar)
data(wine)

# Create learning and test samples
n <- nrow(wine)
sub <- 10:20
X.quanti.train <- wine[sub, 3:29]
X.quali.train <- wine[sub, 1:2]

# Cluster variables using training data
part <- kmeansvar(X.quanti.train, X.quali.train, init = 5)

# Prepare test data
X.quanti.test <- wine[-sub, 3:29]
X.quali.test <- wine[-sub, 1:2]

# Predict scores for new observations
new_scores <- predict(part, X.quanti.test, X.quali.test)

head(new_scores)
```

### Response
#### Success Response (200)
- **scores** (matrix) - Scores for new observations on the synthetic variables.

#### Response Example
```r
# head(new_scores)
#         cluster1  cluster2  cluster3  cluster4  cluster5
# 2H2001  0.543210 -0.234567  0.876543 -0.432109  0.234567
# ...
```
```

--------------------------------

### Predict Scores for New Data

Source: https://context7.com/chavent/clustofvar/llms.txt

Project new observations onto existing cluster synthetic variables using learned coefficients.

```r
library(ClustOfVar)
data(wine)

n <- nrow(wine)
sub <- 10:20
X.quanti.train <- wine[sub, 3:29]
X.quali.train <- wine[sub, 1:2]

part <- kmeansvar(X.quanti.train, X.quali.train, init = 5)

X.quanti.test <- wine[-sub, 3:29]
X.quali.test <- wine[-sub, 1:2]

new_scores <- predict(part, X.quanti.test, X.quali.test)

head(new_scores)
```

--------------------------------

### mixedVarSim - Calculate Similarity Between Variables

Source: https://context7.com/chavent/clustofvar/llms.txt

Computes the similarity between two variables of any type (quantitative-quantitative, qualitative-qualitative, or mixed). Similarity is defined as a squared cosine measure appropriate for each variable type combination.

```APIDOC
## mixedVarSim - Calculate Similarity Between Variables

### Description
Computes the similarity between two variables of any type (quantitative-quantitative, qualitative-qualitative, or mixed). Similarity is defined as a squared cosine measure appropriate for each variable type combination.

### Method
`mixedVarSim(x, y, ...)`

### Parameters
#### Path Parameters
None

#### Query Parameters
- **x** (vector) - The first variable.
- **y** (vector) - The second variable.

### Request Example
```r
library(ClustOfVar)
data(wine)

# Split mixed data
X.quanti <- PCAmixdata::splitmix(wine)$X.quanti
X.quali <- PCAmixdata::splitmix(wine)$X.quali

# Similarity between two quantitative variables
sim_quant <- mixedVarSim(X.quanti[, 1], X.quanti[, 2])
cat("Quantitative-Quantitative similarity:", sim_quant, "\n")

# Similarity between two qualitative variables
sim_qual <- mixedVarSim(X.quali[, 1], X.quali[, 2])
cat("Qualitative-Qualitative similarity:", sim_qual, "\n")

# Similarity between quantitative and qualitative variable
sim_mixed <- mixedVarSim(X.quanti[, 1], X.quali[, 1])
cat("Mixed similarity:", sim_mixed, "\n")
```

### Response
#### Success Response (200)
- **similarity** (numeric) - The calculated similarity value.

#### Response Example
```r
# Output is printed to console
```
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.