Web-based Cluster Engine

Alexander Chen, Rahul Gopalkrishnan, Ben Aneesh, Shamayel Ahmed

Date: April 4, 2013


This project was inspired by Professor Brendan Frey at the University of Toronto. In the machine learning community, people often need modular algorithm architecture for scientist to test out their learning algorithm. The objective of this project is to provide an easy user interface for scientist or people who is interested in data science to cluster their data.


The web interface consist of three default 2D dataset that allows the user to test out the characteristic of different clustering algorithms (k-means clustering algorithm, affinity propagation, spectral clustering, greedy clustering, message-passing clustering). In addition, we implemented the data upload functionality allowing user to upload and cluster their own (.mat)data file.

Web Interface & 2D Dataset

The web interface contains two menu that let the user select the prefered algorithm, dataset and corresponding parameters. As for the 2D dataset, we provide the moon dataset, circle dataset and the blob dataset.

Backend Algorithms

In the back-end server, we provide different algorithms to handle the selected dataset. We offer k-means clustering, affinity propagation, spectral clustering, greedy clustering, message-passing clustering. Below is what it looks like when affinity propagation is executed on the circle dataset and moon dataset.

Data Upload

We also provided data upload functionality. User are able to upload their data in a .mat format which contain a matrix of data. Our system will cluster all the points this matrix and return the label results of the data points.

Like what you read? Give Alex Chen a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.