Machine Learning Clustering Tool

This tool performs clustering on high-dimensional vector data in less time than existing tools. To run this application, installation of .NET Desktop Runtime version 7.0.7 or later is required. If prompted at startup, please follow the instructions to install.

The application is designed to cluster high-dimensional vector data used by LLMs and similar technologies. It enables clustering of vector data that was difficult with existing tools.

Key features of the clustering in this application include: - The maximum number of clusters is 10,000. You can specify the maximum, minimum, and interval for the cluster count. Therefore, it is possible to output clusters in increments of 500 from 5000 to 1000. - The algorithm is not stable, so results may vary with each run. - An option to minimize density fluctuations in vector data is enabled by default, ensuring that cluster size differences remain small even with fluctuations. - Supports multi-threading. Set the appropriate value according to your operating environment.

Limitations of this application include: - It consumes a relatively large amount of memory, a problem related to data handling and a current limitation. - If memory consumption increases, the application's responsiveness may decrease, a problem difficult to mitigate on the app side.

The application is currently free but will require a subscription in the future. It will remain free for academic institutions after the subscription version is released. Please contact us for more information.

If you need the clustering engine used in this application, please contact us.

We continue to release tools related to AI and Machine Learning. If you are interested, we would appreciate your evaluation of them.