RemixAutoML (version 0.5.0)

AutoClustering: AutoClustering

Description

AutoClustering adds a column to your original data with a cluster number identifier. You can run request an autoencoder to be built to reduce the dimensionality of your data before running the clusering algo.

Usage

AutoClustering(
  data,
  FeatureColumns = NULL,
  ModelID = "TestModel",
  SavePath = NULL,
  NThreads = 8,
  MaxMemory = "28G",
  MaxClusters = 50,
  ClusterMetric = "totss",
  RunDimReduction = TRUE,
  ShrinkRate = (sqrt(5) - 1)/2,
  Epochs = 5L,
  L2_Reg = 0.1,
  ElasticAveraging = TRUE,
  ElasticAveragingMovingRate = 0.9,
  ElasticAveragingRegularization = 0.001
)

Arguments

data

is the source time series data.table

FeatureColumns

Independent variables

ModelID

For naming the files to save

SavePath

Directory path for saving models

NThreads

set based on number of threads your machine has available

MaxMemory

set based on the amount of memory your machine has available

MaxClusters

number of factors to test out in k-means to find the optimal number

ClusterMetric

pick the metric to identify top model in grid tune c("totss","betweenss","withinss")

RunDimReduction

If TRUE, an autoencoder will be built to reduce the feature space. Otherwise, all features in FeatureColumns will be used for clustering

ShrinkRate

Node shrink rate for H2OAutoencoder. See that function for details.

Epochs

For the autoencoder

L2_Reg

For the autoencoder

ElasticAveraging

For the autoencoder

ElasticAveragingMovingRate

For the autoencoder

ElasticAveragingRegularization

For the autoencoder

Value

Original data.table with added column with cluster number identifier

See Also

Other Unsupervised Learning: AutoClusteringScoring(), GenTSAnomVars(), H2OIsolationForestScoring(), H2OIsolationForest(), ResidualOutliers()

Examples

# NOT RUN {
#########################
# Training Setup
#########################

# Create fake data
data <- RemixAutoML::FakeDataGenerator(
  Correlation = 0.85,
  N = 1000,
  ID = 2,
  ZIP = 0,
  AddDate = TRUE,
  Classification = FALSE,
  MultiClass = FALSE)

# Run function
data <- RemixAutoML::AutoClustering(
  data,
  FeatureColumns = names(data)[2:(ncol(data)-1)],
  ModelID = "TestModel",
  SavePath = getwd(),
  NThreads = 8,
  MaxMemory = "28G",
  MaxClusters = 50,
  ClusterMetric = "totss",
  RunDimReduction = TRUE,
  ShrinkRate = (sqrt(5) - 1) / 2,
  Epochs = 5L,
  L2_Reg = 0.10,
  ElasticAveraging = TRUE,
  ElasticAveragingMovingRate = 0.90,
  ElasticAveragingRegularization = 0.001)

#########################
# Scoring Setup
#########################

Sys.sleep(10)

# Create fake data
data <- RemixAutoML::FakeDataGenerator(
  Correlation = 0.85,
  N = 1000,
  ID = 2,
  ZIP = 0,
  AddDate = TRUE,
  Classification = FALSE,
  MultiClass = FALSE)

# Run function
data <- RemixAutoML::AutoClusteringScoring(
  data,
  FeatureColumns = names(data)[2:(ncol(data)-1)],
  ModelID = "TestModel",
  SavePath = getwd(),
  NThreads = 8,
  MaxMemory = "28G",
  DimReduction = TRUE)
# }
# NOT RUN {
# }