Seiki Ubukata (生方 誠希)
Osaka Metropolitan University
Department of Core Informatics, Graduate School of Informatics
Department of Information Science, School of Engineering
Associate Professor
E-mail: ubukata (at) omu.ac.jp
1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan
Phone: +81-72-254-9355
1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan
Phone: +81-72-254-9355
Researcher profiles
Software
UbuKit
A clustering and self-organizing map library for Python and JavaScript. Besides k-means, fuzzy and rough clustering and SOM, it provides evaluation metrics and parameter search with TPE.
View on GitHub →
SOM-OLP
A Python implementation of a self-organizing map (SOM) that jointly learns the latent positions of data and the reference vectors. It includes a minimal implementation based on the paper and examples of training and visualization.
View on GitHub →Learning demos
These interactive demos are currently available in Japanese only.
Seeing data as points
An introductory demo on looking at data through points, point clouds and distances.
Open demo (Japanese) →
Learning the centroid
Move representative points to learn about centroids and the idea behind k-means.
Open demo (Japanese) →
K-means demo
Try grouping points into clusters step by step.
Open demo (Japanese) →