Datasets

All 12 datasets used in Categorical Data Clustering via Value Order Estimated Distance Metric Learning can be found in the Data folder. Additionally, six supplementary datasets are included in the Supp_data folder.

Code information

All code is written in Matlab 2022a. Each section serves the following purpose:

• The Metrics folder contains code for evaluating clustering performance.
• The OCL_alg folder details the code of the OCL algorithm.

The script information in the OCL_alg file are as follows:

• The initialization.m and order_initial.m files initialize Q and order, respectively.
• The OCL_main.m file is the main component of the OCL algorithm and executes the outer loop.
• The INOCL.m file implements the inner loop function of the OCL algorithm.
• The Order_choose.m file and the Order_reset.m file are responsible for learning the order and modifying the order of the original dataset, respectively.

Finally, the Execute_Clustering.m file executes the entire OCL algorithm on the provided dataset.

How to run

• Ensure that the OCL_Source_Code folder has been added to the execution path by right-clicking the folder, selecting Add to Path, and then left-clicking Selected Folders and Subfolders.
• Run the Execute_Clustering.m file, and the results will be displayed in the command line window.