ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Unsupervised random forests | JoVE Visualize

Home
Research Domains
Information And Computing Sciences

Unsupervised random forests

Alejandro Mantero¹, Hemant Ishwaran¹

¹Division of Biostatistics, University of Miami, Miami, Florida, USA.

Statistical Analysis and Data Mining|April 9, 2021

View abstract on PubMed

Summary

sidClustering, a novel random forests algorithm, identifies clusters using staggered interaction data and multivariate random forests. This unsupervised machine learning method effectively handles both categorical and continuous variables for robust cluster analysis.

Area of Science:

Machine Learning
Bioinformatics
Data Mining

Background:

Unsupervised machine learning algorithms are crucial for identifying patterns in complex datasets.
Existing methods may struggle with mixed data types (categorical and continuous).
Random forests offer advantages in handling high-dimensional data.

Purpose of the Study:

Introduce sidClustering, a new unsupervised machine learning algorithm based on random forests.
To demonstrate the effectiveness of sidClustering in identifying clusters from mixed data types.
To showcase the algorithm's application in real-world scenarios.

Main Methods:

sidClustering preprocesses features through 'sidification' creating staggered interaction data (SID) main and interaction features.
A multivariate random forest model predicts SID main features.
Multivariate impurity splitting is employed to identify clusters.

Main Results:

The sidification process is unique and reproducible.
Multivariate impurity splitting effectively identifies clusters.
sidClustering successfully identifies clusters from both categorical and continuous variables.

Conclusions:

sidClustering is an adept unsupervised machine learning algorithm for cluster identification.
The method retains the advantages of random forests while handling mixed data types.
Demonstrated utility in simulated data, esophageal cancer studies, and cardiovascular patient charge analysis.

Keywords:

Artificial Intelligence

Fuzzy Computation

Unsupervised Random Forests

staggered interaction data

unsupervised learning