Buildings.Utilities.Clustering.Validation
Validation models for clustering functions.
Information
This package contains models for validation of clustering functions. The examples plot various outputs, which have been verified against comparative results. These model outputs are stored as reference data to allow continuous validation whenever models in the library change.
Extends from Modelica.Icons.ExamplesPackage (Icon for packages containing runnable examples).
Package Content
| Name | Description |
|---|---|
 KMeans_1d
|
Model that verifies the k-means clustering function for 1d data |
| KMeans_2d
|
Model that verifies the k-means clustering function for 2d data |
Buildings.Utilities.Clustering.Validation.KMeans_1d
Model that verifies the k-means clustering function for 1d data
Information
This example tests the k-means clustering algorithm on 1d data.
Extends from Modelica.Icons.Example (Icon for runnable examples).
Parameters
| Type | Name | Default | Description |
|---|---|---|---|
| Integer | n_clusters | 3 | Number of clusters to be generated |
| Real | data[:, :] | [1; 2; 2; 3; 4] | Test data to be clustered |
| Integer | nDat | size(data, 1) | Number of samples |
| Integer | nDim | size(data, 2) | Number of features |
| Real | centroids[n_clusters, nDim] | Centroids of the clusters | |
| Integer | labels[nDat] | Cluster label associated with each data point | |
| Integer | cluster_size[n_clusters] | Size of the clusters | |
| Integer | labelsExp[nDat] | {2,2,2,1,3} | Expected cluster labels |
Modelica definition
model KMeans_1d "Model that verifies the k-means clustering function for 1d data"
extends Modelica.Icons.Example;
parameter Integer n_clusters = 3 "Number of clusters to be generated";
parameter Real data[:,:] = [1;2;2;3;4] "Test data to be clustered";
parameter Integer nDat = size(data,1) "Number of samples";
parameter Integer nDim = size(data,2) "Number of features";
parameter Real[n_clusters,nDim] centroids(each fixed=false)
"Centroids of the clusters";
parameter Integer[nDat] labels(each fixed=false)
"Cluster label associated with each data point";
parameter Integer[n_clusters] cluster_size(each fixed=false)
"Size of the clusters";
parameter Integer labelsExp[nDat]=
{2,2,2,1,3}
"Expected cluster labels";
// Comparison result
Boolean cmp "Comparison result";
initial equation
(centroids, labels, cluster_size) = Buildings.Utilities.Clustering.KMeans(
data, n_clusters, nDat, nDim);
equation
cmp = Modelica.Math.Vectors.isEqual(labels, labelsExp);
end KMeans_1d;
Buildings.Utilities.Clustering.Validation.KMeans_2d
Model that verifies the k-means clustering function for 2d data
Information
This example tests the k-means clustering algorithm on 2d data.
Extends from Modelica.Icons.Example (Icon for runnable examples).
Parameters
| Type | Name | Default | Description |
|---|---|---|---|
| Integer | n_clusters | 3 | Number of clusters to be generated |
| Real | data[:, :] | [1, 1; 1, 2; 2, 1; 2, 1.2; 3... | Test data to be clustered |
| Integer | nDat | size(data, 1) | Number of samples |
| Integer | nDim | size(data, 2) | Number of features |
| Real | centroids[n_clusters, nDim] | Centroids of the clusters | |
| Integer | labels[nDat] | Cluster label associated with each data point | |
| Integer | cluster_size[n_clusters] | Size of the clusters | |
| Integer | labelsExp[nDat] | {2,2,1,2,1,3} | Expected cluster labels |
Modelica definition
model KMeans_2d "Model that verifies the k-means clustering function for 2d data"
extends Modelica.Icons.Example;
parameter Integer n_clusters = 3 "Number of clusters to be generated";
parameter Real data[:,:]=
[ 1, 1;
1, 2;
2, 1;
2, 1.2;
3, 0;
4, 0]
"Test data to be clustered";
parameter Integer nDat = size(data,1) "Number of samples";
parameter Integer nDim = size(data,2) "Number of features";
parameter Real[n_clusters,nDim] centroids(each fixed=false)
"Centroids of the clusters";
parameter Integer[nDat] labels(each fixed=false)
"Cluster label associated with each data point";
parameter Integer[n_clusters] cluster_size(each fixed=false)
"Size of the clusters";
parameter Integer labelsExp[nDat]=
{2,2,1,2,1,3}
"Expected cluster labels";
// Comparison result
Boolean cmp "Comparison result";
initial equation
(centroids, labels, cluster_size) = Buildings.Utilities.Clustering.KMeans(
data, n_clusters, nDat, nDim);
equation
cmp = Modelica.Math.Vectors.isEqual(labels, labelsExp);
end KMeans_2d;